Affective Events Theory A Theoretical Discussion o PDF

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/228079335
Affective Events Theory: A Theoretical Discussion of The Structure, Cause and

Consequences of Affective Experiences at Work
Article in Research in Organizational Behavior · January 1996
CITATIONS READS
1,264 13,047
2 authors, including:
Russell Cropanzano
University of Colorado Boulder
143 PUBLICATIONS 18,687 CITATIONS
SEE PROFILE
All content following this page was uploaded by Russell Cropanzano on 13 October 2017.
The user has requested enhancement of the downloaded file.

! AFFECTIVE EVENTS THEORY:
A THEORETICAL DISCUSSION OF THE
i
STRUCTURE, CAUSES AND CONSEQUENCES
I OF AFFECTIVE EXPERIENCES AT WORK
Howard M. Weiss and Russell Cropanzano
ABSTRACT
In spite of accepted definitions of job satisfaction as "affect" very little is known
about the causes and consequences of true affective experiences in work settings.
Working from the basic literature on moods and emotions, we introduce a theory
of affective experience at work which emphasizes the role of work events as
proximal causes of affective reactions. We discuss the structure of affective
experiences, their situational and dispositional causes and their effects on
performance and job satisfaction.
Research in Organizational Behavior, Volume 18, pages 1-74.

!
Copyright @ 1996 by JAI Press Inc.
All rights of reproduction in any form reserved.
ISBN: 1-55938-938-9
2 HOWARD M. WEISS and RUSSELL CROPANZANO
Cranny, Smith, and Stone (1992) define job satisfaction as "an affective (that
is, emotional) reaction to a job that results from the incumbent's comparison
of actual outcomes with those that are desired (expected, deserved, and so on)"
(p. 1). This is a rather curious definition of a construct, including as it does
both the essential variable (affective reaction) and its presumed causes
(outcome-standard comparisons). The fact that there is general agreement with
this definition (their conclusion) makes the inclusion of a theory of the
construct's causes in the definition of the construct no less troublesome.
However, confusion of construct and causes is not the most troublesome
aspect of this definition. More difficult is the fact that job satisfaction is not
an affective or emotional reaction to a job, or at least it is not as it typically
has been studied and operationalized. Judging from the nature of most job
satisfaction questionnaires it would be more accurate to argue that job
satisfaction is a positive or negative evaluative judgment of one's job or job
situation. This is decidedly not the same thing as an affective or emotional
reaction.
What then is the connection between affect and satisfaction? The answer
to this question is found by clearly distinguishing cause and effect. Satisfaction
is an evaluative judgment about one's job that partly, but not entirely, results
from emotional experiences at work. It also partly results from more abstract
beliefs about one's job. Together, affective experiences and belief structures
result in the evaluation we call job satisfaction.
Treating job satisfaction as a summary evaluation with both affective and
belief antecedents is quite consistent with current positions on attitude
formation. While it has long been recognized that attitudes have affective and
cognitive correlates, the real meaning of this point, as well as research on the
interplay of the two components, is only now taking shape. Breckler and
Wiggens (1989) provide a very cogent discussion of this issue. They argue that
attitudes are general evaluations encompassing both affect and cognition. The
affective component refers to feelings that are "engendered" by the attitude
object and "represents emotional experience associated with the attitudinal
object."The cognitive component refers to beliefs about the object, the location
of the object on dimensions of judgment. Overall evaluations are influenced
by both of these components to greater or lesser degrees.
Making this distinction between satisfaction as judgment and the affective
experiences which influence this judgment leads quickly to the conclusion that
satisfaction and affective experience should be treated as separate phenomena
with distinct but overlapping causes and consequences. This, in turn, suggests
the importance of studying affective reactions at work independent of job
satisfaction.
Paradoxically, in spite of the "accepted" definition of satisfaction as affect
or emotion, we know precious little about emotional reactions at work. In this
paper we intend to describe a theoretical position on affective reactions which
Affective Events Theory 3
we hope will serve as a guide to research on the topic. We will suggest a structure
for emotional reactions as well as ideas about their causes and consequences.
One of the consequences is job satisfaction or dissatisfaction and thus one could
argue that a piece of the total position constitutes a piece of a theory of job
satisfaction. Theory is perhaps too pretentious. The term is meant only to
suggest that there will be conceptual variables integrated into a general
framework. In any case, this paper is offered in the spirit of William James
(1890) who said "At a certain stage in the development of any science a degree
of vagueness is what exists with profit."
THREE THEORETICAL APPROACHES

TO JOB SATISFACTION
While the position we will outline focuses on the structure, causes, and
consequences of affective experiences, it rests squarely in the overall tradition
of research on job satisfaction. Therefore, we begin with a general overview
of the previous positions on job satisfaction, and then contrast these positions
with one that focuses on affective experiences.
Most specific theoretical positions on job satisfaction can be taken as variants
of three general approaches, which we refer to as the "Cognitive Judgment
Approach," the "Social Influence Approach," and the "Dispositional
Approach." In discussing elements of these positions we will avoid any
evaluation of the research literature as good reviews exist elsewhere (Arvey,
Carter, & Buerkley, 1991; Locke, 1976).
Cognitive Judgment Approach
By far, the cognitive judgment approach has dominated the theoretical

landscape. It is represented most generally by what Lawler (1973) referred to
as Discrepancy Theory, and in specific forms, for example, by Katzell's (1964)
and Locke's (1976) outcome-value discrepancies, by Ilgen's (197 1) discussion
of outcome-expectancy discrepancies and by Porter's (1962) position on
outcome-need discrepancies. It is also represented by Lofquist and Dawes'
(1969) Theory of Work Adjustment, by theories of job satisfaction within the
expectancy-value-instrumentality tradition (Mitchell, 1974) and by equity/
justice theories of satisfaction (Greenberg, 1982). Lest anyone believe that the
advanced age of these seminal references suggests that the position is no longer
considered useful, recent papers from these positions continue to appear
(Dawes, 1992; Konovsky, Folger, & Cropanzano, 1987; Rice, McFarlin, &
Bennett, 1989; Stone, 1992).
While each of the theories has its particular constructs, a general structure
exists for all theories in this tradition (see Figure 1). In this general structure
the work environment is represented as a set of concrete or abstract features

Cjob characteristics, pay levels, promotion opportunities, etc.). These features
are perceived, not always accurately, by job incumbents who compare their
perceptions to some set of standards (values, needs, etc.). In some versions the
features have meaning only in a work context (e.g., pay levels, career
opportunities) while in other versions more abstract psychological properties
are used (e.g., autonomy). In any case, some sort of arithmetic function
(differences, weighted differences, ratios, etc.) is used to assess the match
between perceptions and standards and this match is the proximal cause of
job satisfaction. Over the years the nature of the standards has been a major
point of disagreement (see Locke, 1976, for the classic discussion) but the
fundamental and underlying structure of the cognitive judgment approach is
consistent across different theories in this tradition.
The overarching domination of this approach suggests that it might be useful
to examine some of its assumptions. For the most part, these assumptions turn
out not to be logical necessities. Rather, they are simply characteristic of
research and writing from this tradition.
To begin with, implicit in the cognitive judgment approach is the relative
stability of job satisfaction and, consequently, the general unimportance of time
as a factor in the study of work attitudes. Certainly environmental features
can change and so too can standards, but by and large these constructs are
treated as being relatively stable on a day-to-day basis. As such, it makes little
difference whether we measure satisfaction on Monday or Wednesday, July
or September. Clearly, the cognitive judgment approach would not suggest
the usefulness of measuring affective reactions on a daily basis or treating
fluctuations as anything but error.
Features
Figure 1. Cognitive Judgment Approach

This point is ironic because much has been made recently of the stability
of work attitudes and the implication of that stability for dispositional versus
environmental explanations (Gerhart, 1987; Staw & Ross, 1985). The basic
conclusion that any demonstrated stability is compatible with both positions
(Davis-Blake & Pfeffer, 1989; Gerhart, 1987) is exactly what one would expect
because in the cognitive judgment approach both the person and the
environment are described in "dispositional" terms. Characteristics of people
interact with characteristics of the environment.
Within the cognitive judgment approach a dimensional structure for
satisfaction has evolved which focuses on the attributes or features of the
environment. This is the distinction generally made between overall and facet
satisfaction and it is a natural result of a feature evaluation model of work
attitudes. On the other hand, with one very notable exception (see Herzberg,
Mausner, & Snyderman, 1959),satisfaction itself is treated as a unitary concept.
We may ask, "How satisfied are you with" this or that but the experience of
satisfaction is not broken down any further.
This is reasonable for studying satisfaction, the overall evaluative judgment
about one's job. However, if it is affect we are after then we must recognize
that affective reactions have their own phenomenal structure. For example,
a person may be angry, frustrated, sad, or ashamed. All are negative affective
reactions yet they may result in quite different behavioral consequences. The
possibility of meaningfully distinct affective experiences at work has been
ignored by researchers accepting job satisfaction as a measure of affect.
In addition, the focus on feature evaluation has led to a simultaneous neglect
of the study of what actually happens at work. A pat on the back, receiving
a bonus, an award at a dinner are the real experiences that somehow influence
responses to the item "my job provides recognition." Yet, we know very little
about these experiences, how people react to them, and how they affect overall
evaluations.
Finally, and in spite of the definition of satisfaction as emotion, one could
honestly ask "where is the emotion in the study of job satisfaction?" How do
we account for the reservationist in Studs Terkel's (1974) Working who says
about her job "I hated it with a passion. Getting up in the morning, going
to work feeling, Oh my God, I've got to go to work." Certainly not by asking
her to tell us how much of some feature she wants and how much of some
feature she thinks she gets.
Social Influences Approach
In the late 1970s the Social Influences approach, most notably in the form
of Social Information Processing Theory (SIP) (Salancik & Pfeffer, 1977,
1978), was presented to the field. An excellent review of this position is provided
by Zalesny and Ford (1990).
The basic idea of SIP is that the social environment has both direct and
indirect influences on judgments about work. It has direct influences on overall
attitudes (Adler, Skov, & Salvemeni, 1985) as well as indirect influences on
the perceptions (Weiss & Shaw, 1979) and standards (Weiss, 1977) which feed
into attitude judgments.
Over the years, many authors have taken great pains to generate a
controversy over whether Social Information Processing is a valid alternative
to the traditional cognitive judgment approach (see, most recently, Stone,
1992). At its core, with regard to most of SIP'S elements, there is nothing
inherently contradictory about the SIP and traditional Cognitive Judgment
approaches (see Figure 2). Social Information Processing is more a
complement than an alternative to Cognitive Judgment in that it fills in some
loose ends about factors which influence the basic variables in the judgment
process. That is, the Cognitive Judgment approach argues that we make
evaluative job satisfaction judgments based on our perceptions of whether or
not some desired standards are being met. The Social Influences approach
maintains this cognitive process. It simply adds the caveat that social
information is a major input into our perceptions and standards.
What then is the source of the idea that SIP and Cognitive Judgment are
antagonistic positions? In our opinion, part of the problem is the loose language
of the original Salancik and Pfeffer articles which introduced new and
unnecessary constructs that were difficult to understand and operationalize.
In addition, part of the problem is their presenting "need fulfillment" as
representative of all discrepancy positions (thus criticizing all by criticizing the
concept of needs).
Features
Figure 2. Cognitive Judgment and Social Influences Approaches

However, we also believe that there is a genuine source of conflict between

the positions and it centers on one "controversial" aspect of Salancik and
Pfeffer's (1977) initial position. That aspect is the frequent reference to the
"Social Construction of Reality," as if attitudes were continuously being
I constructed and reconstructed in a social context. This, we would argue,
suggested to some readers that job attitudes have an ephemeral, artificial
quality, independent of the reality of an external world of job and work
characteristics. The importance of this point is illustrated by this question:
"With regard to the construction of job attitudes, is this construction/judgment
I
made and stored and then recalled when attitudes are requested, or is the
construction/judgment made and remade each time the attitude is requested,
with potentially different information used on each occasion?" Traditional
Cognitive Judgment theorists would take the former position, as indicated
earlier, although they would not suggest that these attitudes could not be revised
in the face of new information. Social Information Processing, some might
argue (Zalesney & Ford, 1990), suggests the later position. To a great extent
this element of Salancik and Pfeffer's position, an element frankly not well
I articulated in the initial papers nor ever empirically explored in the literature,
does not sit well with those advocating the traditional approach. It has led
to a broad brush painting of the whole SIP position as somehow antagonistic
to the Cognitive Judgment approach.
We will not try to resolve this controversy. We will only say that current
1 thinking on attitude expression is giving more and more weight to the idea
that attitudes generally are not stored and recalled but are "c,onstructed" on
demand (Wilson & Hodges, 1992). Further, each construction involves the
integration of both stored information relevant to the attitude object and
contextual information unrelated to the attitude object. Social information can
easily be understood as important contextual information.
Dispositional Approach
The basic idea of the Dispositional approach is this: to some degree, a

person's job satisfaction reflects his or her general tendency to feel good or
bad about all aspects of life and this general tendency is independent of the
specific nature of the job, its positive or negative features. Recently, Judge
\
,
(1992) has provided an excellent review of the Dispositional position and so
we will only touch on the highlights, again focusing more on the nature of
the approach than on any evaluation of the research results.
Joseph Weitz first discussed a dispositional approach to job satisfaction and
developed what he called the gripe scale to measure it (Weitz, 1952), but current
t interest on the topic has been stimulated by the work of Staw and Ross (1985).
Staw and Ross argued that previous theories of job satisfaction had
overemphasized situational causes at the expense of dispositional ones. To find
support for their position they examined the stability of job attitudes over a
five-year period for a national sample of men. Their sample did in fact show
attitudinal stability, as judged by the correlations of attitudes measured in
different years. In spite of the fact that these correlations were smaller among
those men who had changed jobs or careers than among those who did not,
their data can be taken as providing support for some stability in job attitudes.
As has been noted by a number of people (Davis-Blake & Pfeffer, 1989;
Gerhart, 1987), stability by itself does not provide particularly strong evidence
for dispositional factors, as the stability could just as easily be a function of
stable job features as dispositions. More compelling would be the identification
of the relevant individual differences variables underlying the dispositional
component. In'fact, Staw and Ross (1985) speculated on the nature of the
underlying personality variables suggesting that individual differences in
affective tendencies, perhaps biologically based, might be responsible.
Since Staw and Ross (1985), research on the dispositional component of
job satisfaction has followed two paths. One path has tried to explicate the
particular personality traits responsible for the dispositional nature of
satisfaction. The other path is embodied in the work of Arvey, Bouchard, Segal,
and Abraham (1989) who have attempted to demonstrate an inherited tendency
to report satisfaction/dissatisfaction.
Personality Determinants of Dispositional Satisfaction
Almost all of the recent research on personality and job satisfaction has
looked at the two personality traits of Positive Affectivity and Negative
Affectivity. These are personality traits which predict general emotional
tendencies in people. People who are high on Positive Affectivity (PA) tend
to be lively, sociable, and often in a positive mood. People who are high on
Negative Affectivity (NA) tend to be more distressed and unhappy, focusing
on the negative side of things.
Because job satisfaction, like all attitudes, has affective as well as belief
components, it is not surprising that differences in affective tendencies have
been shown to be associated with differences in job satisfaction. For example,
Levin and Stokes (1989) conducted a laboratory study in which they asked
people who were either high or low on NA to work on interesting or boring
tasks. They found that quite independent of the type of task, their satisfaction
with the task was influenced by their degree of Negative Affectivity. Similarly,
in two field studies Cropanzano, James, and Konovsky (1993) showed that
both NA and PA correlated with global satisfaction. Recently, Watson and
Slack (1993) found that PA predicted job satisfaction as long as two years later.
Clearly, general affective tendencies are partly responsible for dispositional
aspects of job satisfaction, but what is the significance of this finding? In our
opinion, simply knowing that some personality variables account for more
variance in satisfaction is not by itself very interesting. Accounting for more

variance is much less important than building a theoretical framework with
both dispositions and situations existing harmoniously in the service of
explanation. Ultimately, the disposition has to enlighten underlying process.
Do these findings on affective dispositional correlates of job satisfaction help
to expand our conceptual understanding of this construct? We think they do
in an important but limited way. An underlying theme in this research is the
significance of affect. Unlike early correlational approaches to satisfaction,
current dispositional studies have been focused on affective tendencies. It
appears that these dispositions are capturing the affective, as opposed to the
belief, component of job attitudes (Brief & Roberson, 1989). This in turn
suggests that the effect of these dispositions is mediated by affective processes.
For example, George (1989) and Weiss, Nicholas, and Daus (1993) suggest
that emotional dispositions like Negative and Positive Affectivity influence
mood states at work and these mood states can influence satisfaction as well
as other important work behaviors. The findings on affective dispositions
remind us that there are affective influences on satisfaction judgments. It
remains to more fully explicate those affective based processes.
Genetic influences on Job Satisfaction
Staw and Ross (1985) suggested that there may be biologically based
explanations for their dispositional findings. Arvey, Bouchard, Segal,
Abraham (1989) took this suggestion a step further and examined the
possibility of a genetic component to satisfaction by way of estimating its
hereditability, hereditability being the proportion of phenotypic variance
accounted for by genetic factors (Willerman, 1979).
In the Arvey, Bouchard, Segal, and Abraham (1989) study, 34 pairs of
monozygotic twins reared apart (MZA) were administered the Minnesota
Satisfaction Questionnaire and the intraclass correlation for the twins'
satisfaction levels was taken as an estimate of hereditability. Using this method,
Arvey et al. (1989) estimated an hereditability level of .31 for overall
satisfaction. They also concluded that the hereditability of intrinsic facets is
stronger (.32) than is the hereditability of extrinsic factors (. 1 I). Furthermore,
partialling out the effects of similarity in job types had minimal effects on these
hereditability estimates, leading Arvey et al. to conclude that self-selection into
similar jobs could not account for the satisfaction results.
Arvey, McCall, Bouchard, and Taubman (1994) reported a replication of
the earlier results, this time using both monozygotic and dizygotic twins, not
necessarily reared apart. In this type of study, hereditability is estimated by
comparing the correlations for the MZ twins, who share all genetic
components, with the correlations for DZ twins, who on average share only
50% of genetic components. Here Arvey et al. (1994) found results consistent
with the earlier findings but with generally smaller hereditablity estimates (less
than 20%).
In spite of methodological criticisms raised by one of us (Cropanzano &
James, 1990) our belief is that Arvey's results suggest that there is some genetic
influence on satisfaction (see also Bouchard, Arvey, Keller, & Segal, 1992).
However, that effect is apparently small. Even using Arvey's initial estimate,
the hereditability is not higher than .30 and, given some of the methodological
issues raised by Cropanzano and James and also the results from Arvey's own
replication, an herititability somewhat lower than .30 is probable.
As it is with personality correlates, simply knowing that there is a genetic
component to satisfaction is not the same thing as understanding it. For us,
the key question with regard to hereditability, as with dispositions generally,
is "does this finding enlighten us about the processes involved in job
satisfaction?" By themselves, these hereditability findings do not offer much
in the way of psychological explanation. However, they can point in productive
directions, if carefully examined.
One direction, naive in our opinion, would be to search for physiological
process explanations to substitute for ps'ychological ones. This would be naive
given the small hereditability coefficients and the fact that even strong genetic
effects typically define a "range of reaction" further influenced by
environmental factors. Besides, as Plomin (1990) suggests, it seems likely that
genetic effects on behavior are the results of the interactions of many genes
making small interactive contributions. Physiological and psychological
explanations can certainly complement and enrich each other. However, the
history of attempts to baldly substitute physiology for psychology is simply
not impressive.
A more useful direction is to search for psychological processes that can
incorporate these genetic findings and their likely physiological consequences.
What we mean is that future theories can ill afford to ignore these hereditability
findings. They have to be incorporated into general theories of job satisfaction,
but not as boxes labeled "heredity" with arrows pointing to job satisfaction
but as signposts pointing the way to appropriate proximal, psychological
causes.
AFFECTIVE EVENTS THEORY: AN OVERVIEW

Our own reading of the "signposts" directs us toward affective reactions and
we will now present an overview of our position, a position we call Affective
Events Theory or AET. Our overview will highlight some of the more
important elements of AET, contrasting our position with more traditional
positions on job satisfaction along the way. More detailed discussions of the
components will constitute the remainder of the paper.
Affective Events Theoy 11
In contrasting Affective Events Theory with traditional theories the first

point of departure is that Affective Events Theory focuses on the structure,
causes and consequences of affective experiences at work. As described earlier,
satisfaction is an evaluative judgment made about one's job. While affective
experiences may influence that judgment, satisfaction and affect are not
equivalent constructs. In its delineation of the factors which influence job
satisfaction, Affective Events Theory focuses on affective experiences as a
counterbalance to theories which exclusively focus on judgment processes.
However, here affective experiences are the more central phenomena of interest
with job satisfaction being one consequence.
As a second point of departure, Affective Events Theory directs attention
away from features of the environment and towards events asproximal causes
of affective reactions. Things happen to people in work settings and people
often react emotionally to these events. These affective experiences have a direct
influence on behaviors and attitudes and the nature of these effects has not
been explored. We are not dismissing the relevance of features but we are
tentatively suggesting that environmental features influence affect primarily by
making affective events (or the recall or imagination of affective events) more
or less likely.
Affective Events Theory also adds time as an important parameter when
examining affect and satisfaction. Research on mood and emotion clearly
indicates that affect levels fluctuate over time and that the patterns of these
fluctuations are predictable to a great extent. We are proposing that these
patterns of affective reactions influence both overall feelings about one's job
and discrete behaviors at work.
Paying attention to patterns of affective experience over time is in direct
contrast to traditional approaches to job satisfaction in which the time of
measurement is given no theoretical importance. In such approaches,
satisfaction, as well as indices of predictors and consequences (performance,
for example) are assessed at some arbitrary point in time. Ignoring time is
consistent with a theoretical position which focuses on the effects of
environmental features because such features are considered relatively stable.
Unfortunately, this approach ignores the importance of affective variation, a
mistake which may contribute to the failure to find affect-performance
relationships of any substance.
Finally, Affective Events Theory considers the structure of affective reactions
as important as the structure of environments. Dimensional structures of job
satisfaction focus on dimensions of the attitude object. That is, the job is the
object and its dimensions are features like pay or supervision. AET recognizes
that affect itself is multidimensional and emphasizes the importance of the
structure of the psychological experience. People can feel angry, frustrated,
proud or joyful and these different reactions have different behavioral
implications.
I..
HOWARD M. WEISS and RUSSELL CROPANZANO
Work
Environment Judgment
Features Driven
Behaviors
( Affective Reactions ( -
Figure 3. Affective Events Theory: Macro Structure
In Figure 3 we have presented a tentative macro structure to the affective events

position. While we will use the remainder of the paper to describe its components
in more detail, it seemed useful to begin by walking through the basic elements.
At the core of the position are affective experiences. A key question involves
how these experiences are to be described. Moods are different from emotions
and emotions, at least, can be described at different levels of differentiation.
We will address this question later and will suggest that the appropriate
structure depends on the problem being addressed.
Fundamental to the theory is the idea that affect levels fluctuate over time
and that the causes of these patterns of affect can be examined in terms of
endogenous components, such as known cycles in mood or affective
dispositions, and exogenous components, such as affectively relevant events
which constitute shocks to existing patterns. Dispositions can also influence
the way events produce affective reactions. Work environments are seen as
having an indirect influence on affective experience by making certain events,
real or imagined, more or less likely.
The consequences of affective experience are both attitudinal and behavioral.
Affective experiences have a direct influence on job satisfaction. This influence
corresponds to the affective aspect of attitudes. Features have both direct and
indirect influences: directly by evaluation in the "cognitive" judgment part of
satisfaction and indirectly through their influence on the likelihood of various
events. Note, however, that this reference is made to features generally. We
are not suggesting that each feature of the work environment has both direct
and indirect effects.
Afictive Events Theory 13
Finally, behaviors are grouped into two categories: affect driven behaviors
and judgment driven behaviors. Affect driven behaviors follow directly from
affective experiences and are not mediated by overall attitudes. They are
influenced by processes like coping or mood management or by direct effects
of affect on cognitive processing or judgment biases. Judgment driven
behaviors are mediated by satisfaction. They are the consequences of decision
processes where one's evaluation of one's job is part of the decision matrix.
PREVIOUS RESEARCH CONSISTENT WITH AET

While it is obviously tempting to argue that no important research has focused
on the role that affective events play in job satisfaction or that affect has been
ignored entirely, this would be far from the truth. As we indicated earlier,
affective dispositions are currently enjoying some popularity (see Judge, 1992)
and the effects of mood on various aspects of work behavior has received some
attention (see, for example, George, 1989, 1992). These efforts, however, do
not represent attempts to provide coherent theoretical perspectives on the topic.
Historically, we believe, two such endeavors are close to what we offer and
therefore should be discussed in this paper. Both, by and large, have been
forgotten in today's discourse on the topic.
Hersey (1932)
In 1932 Rexford Hersey published his research on Workers' Emotions in

Shop and Home. Hersey's research is now mostly forgotten yet it still stands
as the seminal piece on emotional reactions at work. For one year Hersey
intensively studied the emotions and behaviors of 17 skilled workers in two
departments of a railroad car and locomotive repair facility of a large public
utility. Each man was interviewed four times a day, first for 13 weeks and then
for periods of 10 weeks and 13 weeks after periods of eight and four weeks,
respectively, during which no measurements took place. The interview was
composed of a number of questions including what would today be called a
mood checklist (22 items such as happy, hopeful, tense, angry, etc.). Hersey
was also able to assess daily productivity as well as individual differences in
physical and psychological attributes.
Hersey's general "theoretical" position focused on the concept of life crises.
He argued that life, including work life, is filled with a variety of major and
minor crises which demand adjustment. The effects of these crises could be
seen by examining daily patterns of emotional reactions. While Hersey did not
empirically examine this proposition he did obtain some interesting findings
with regard to mood and behavior. For our purposes, three findings are of
particular interest. First, Hersey was able to chart daily mood levels, scored
as positive and negative, and document daily fluctuations in affect. Second,

by assessing both daily mood levels and daily performance levels he was able
to demonstrate a definite relationship between emotional state at work and
productivity. Interestingly, this relationship was not symmetrical in that the
negative effects of a negative emotional state were much more pronounced than
were the positive effects of a positive state (see Figure 4).
Third, without benefit of modern statistical techniques Hersey was able to
discern definite mood cycles over the course of the year. These cycles varied
across individuals. Different workers had longer or shorter periods to their
cycles.
Hersey's research approximately coincided with the Hawthorne studies and
with Hoppock's research on job satisfaction. For whatever reasons, it is
certainly true that both the Hawthorne and Hoppock studies are mentioned
in almost all current Industrial and Organizational Psychology textbooks while
Hersey is virtually unknown to today's researchers. Yet it is probably true that
Hersey's work could be published intact today (albeit with a little analytical
updating) while neither Hawthorne nor Hoppock would withstand current
review processes.
Herzberg, Mausner, a n d Snyderman (1959)
Herzberg maintained that satisfaction and dissatisfaction were not poles of

a single continuous dimension but separate and distinct dimensions. The field
responded with cries of "methodological artifact," "confusion of agents and
events," and "attributional processes." It seems clear that if a modern day
Herzberg were to arrive on today's scene with the Two Factor Theory in hand
the reactions would be substantially muted. After all, today we have Positive
and Negative Affect (Watson & Tellegen, 1985, as discussed later) and
Herzberg's dimensions of satisfaction would be perfectly consistent with that
particular dimensional organization of mood self-reports.
For our purposes it is useful to note a few relatively unique aspects of
Herzberg's work, apart from his Two Factor conceptualization. To begin with,
Herzberg understood and emphasized the variability of work attitudes. Using
his critical incidents method he asked respondents to "think of a time when
you felt especially good or bad about your job." Nothing in this question
suggested that people respond in terms of different jobs or that they had to
separate these incidents by any particular length of time. In fact, he said that
implicit in his new approach is "the notion that job attitudes varied for each
individual from one period to another" (p. 13) and that these variations could
be linked to variations in diverse criteria.
For Herzberg, the primary causes of these periods of satisfaction or
dissatisfaction were specific work events. "1 felt terrible when I was passed over
for the promotion" (p. 42), "When I finally knew I had the problem licked,
I felt higher than a kite" (p. 42), "A warehouse checker is ordered by his
supervisor to go out in the rain and check a group of freight cars" (p. 21).
Careful review of these incidents indicate that some of the responses described
specific events and others described longer term "situations." Herzberg referred
to them all as events but clearly some respondents answered in terms of shorter
time frames and described specific occurrences on the job and others answered
in terms of longer time frames and described general features of the work
situation. However, even these "long range sequences" as Herzberg referred
to them, had implicit shorter term events at their core. For example, a salesman
describes a long-term situation of frustration but includes in his description
that "the boss was too busy to train him and seemed annoyed whenever (he)
asked questions" (p. 22); or, as another example, an engineer expresses career
frustrations but says that he felt bad because there was a strike going on and
management refused to let him participate in the negotiations (p. 22).
Interestingly, while Herzberg described the proximal causes of satisfaction
as work events, he did not stay at this level of analysis for long. Instead, he
used his events to develop descriptions of environments which facilitated the
events (interesting work, possibility for advancement, possibility for growth,
etc.) and then kept his focus on these environmental features as he developed
his concepts of "motivators" and "hygienes." This focus on features made it
easier to incorporate his findings with a need satisfaction approach to
satisfaction, specifically Maslow's Hierarchy of Needs (Maslow, 1954).
However, a careful reading of Herzberg clearly indicates that in his theory
features operate by making certain events more or less likely.
If we keep the distinctions between events and features and the distinctions
between affective reactions and overall evaIuations in mind, the great
controversy between one factor and two factor theories ofjob satisfaction seems
rather inconsequential. Our overall evaluation of our job is unidimensional,
consistent with a one factor approach. Further, at any given time our current
affective state is either positive or negative (Diener & Emmons, 1984) and the
affective consequences of work events can be either positive or negative.
However, certain kinds of features can be conducive to the occurrence of
positive but not negative events and other kinds of features can be conducive
to negative events but have little influence on the frequency of positive events.
But, again, the construct we refer to as job satisfaction is still a unidimensional
evaluation of one's job.
THE NATURE OF EMOTIONS AND MOODS

At this point we will start to examine the elements of AET in more detail.
We will begin at the core of the theory, with a discussion of the nature of
affective reactions, their definitions and structural representations. This will
be followed, in turn, by discussions of the way events come to elicit affective

reactions, what is known about other antecedents of mood and emotion,
affective cycles, the influence of affective reactions on job satisfaction and,
finally, our ideas about the performance implications of affective reactions.
We should say immediately that the study of emotion has a long history
and the literature is enormous and fragmented (see Plutchik, 1994, for an up-
to-date summary.) In fact, the scientific study of emotion predates the formal
birth of psychology with the writings of Charles Darwin (1872[1965]).Emotion
was there, too, at the birth of psychology, as two of psychology's "fathers,"
William James and Wilhelm Wundt, wrote extensively on the topic.
Our objective cannot be to review all that is known about affect, moods,
and emotions. Rather, we hope to summarize those facts and theories which
have relevance to work experience. We are intentionally selective without trying
to imply that the theories we have chosen to present are the accepted positions
in the field. No theory currently meets this criterion.
The objective we d o acknowledge is to convince the reader of the validity
of the basic structure of our theory, the importance of events, the difference
between affect driven and judgment driven behaviors, and so forth and then
selectively use the current literature on moods and emotions to fill in the details
of the structure while suggesting certain avenues of research. In doing this we
recognize that other organizational researchers might accept the overall
framework but be guided by different positions in filling in the details.
Defining Emotions
Emotions and moods are both affective states yet rarely have psychologists
made explicit attempts to differentiate them. This curious situation is mostly
due to the fact that these two types of affective experiences have two different
research traditions. The research tradition involving emotions is long and
varied. The tradition focusing on moods is shorter and more focused.
Emotions are intuitively well understood yet a definitive definition of
emotion has been difficult to come by. The difficulty in developing a definition
seems to arise from the observation that an emotional reaction is not one
reaction, but a constellation of related reactions. Nonetheless, most definitions
seem to settle on a few essential components. We will use Frijda's (1993)
summary to introduce these components. According to Frijda, every emotional
experience has four main components. To begin with there is the experiential
component of affect, "the irreducible aspect that gives feelings their emotional,
noncognitive character" (p. 383). Next, Frijda argues, a person is generally not
aware of feeling good or bad but rather one is aware of the pleasantness or
unpleasantness of the eliciting event and therefore the experience of affect is
intricately tied to the appraisal of that event. Most definitions of emotions
include the idea that there is an overall affective experience and a consequent
cognitive appraisal process (Plutchik, 1994). Third, for Frijda and for most
theorists, emotions are also characterized by a wide variety of physiological
bodily changes. Finally, Frijda suggests that the experience of an emotion
includes an action readiness, a general readiness to deal with the environment
through increased arousal and vigilance.
Implicit in all definitions is that an emotion is a reaction to an event. It is
not a trait, although there can be trait differences in chronic affect levels or
in reactivity to specific events. Also implicit in all definitions is that emotions
have event or object specificity. As Frijda (1993) says "Emotions have an object,
they are about something.. . One is happy about something, angry at someone,
afraid of something" (p. 381). Finally, even when one acknowledges the
multiple components of the emotional experience (affect, physiology, etc.) it
is the experience that remains fundamental.
Defining Moods
Frijda (1993) and Morris (1989) both tell us that moods are most frequently
distinguished from emotions by three features: intensity, duration, and
diffuseness. More specifically, moods, as compared to emotions, are thought
to be less intense, of longer duration and lack specificity with regard to a
particular object or behavioral response.
Frijda (1993) and Morris (1989) also tell us that the first two criteria are
not very useful. To begin with, moods can vary greatly in their duration. One
often thinks of moods lasting for hours or even days, yet mood manipulations
in the laboratory can last less than a few minutes and still have the effects
attributed to real mood experiences. Emotions, on the other hand, can last
for long periods of time. In addition, diffuse affective states like anxiety or
depression can be of very high intensity while specific emotional reactions can
be rather mild.
Frijda (1993) and Morris (1989) conclude that the real distinguishing feature
between moods and emotions is diffuseness in terms of both object and
response. Emotions are affective states directed at someone or something
(Frijda, 1993). Moods, on the other hand, lack an object to which the affect
is directed. Lazarus (1991a) makes a similar point when he says that moods
are vague and "lack a contextual provocation."
None of these authors are suggesting that moods do not have specific causal
antecedents, only that the phenomenal experience of the mood does not include
the causal factor. The importance of the experiential disconnection between
the affect state and its cause is further emphasized in Frijda's (1993) suggestion
that an emotion turns into a mood when one loses the focus on the precipitating
event or object. Similarly, making the cause of the mood salient may transform
a mood into a weak emotion. This latter position is consistent with research
I
which demonstrates that when people are made aware of the cause of their
mood state many of the global effects of mood are eliminated (Clore, 1992).
k Mood researchers also argue that moods, more than emotions, are diffuse
in terms of elicited responses in that they influence a wide variety of cognitive
and behavioral responses which are not connected to the original source of
the mood (Isen, 1984; Morris, 1989). Obviously, we agree with the argument
that moods have diverse effects but we would caution that this distinction
between mood and emotion can be taken too far. While it is true that the effects
of emotions tend to be more targeted toward dealing with either the source
of the emotion or the emotional state, they too can have generalized behavioral
1 effects mediated by activation or arousal levels. In addition, many of the global
'i effects of moods may result from rather specific attempts to manage one's
subjective state. This being said, it still remains that the effects of mood tend
r to be less dependent on the nature of the cause of the mood. This is consistent
! with the idea that the cause is not part of the phenomenal experience.
I
I The Stucture of Emotions
Researchers on moods and emotions have spent a good amount of time and
energy trying to dimensionalize, classify and categorize these two types of
affective responses. In this section and the next we will present a summary
of these efforts. It is our observation that the dimensionalization or
categorization objectives of emotion researchers appears to be quite different
from the objectives of psychologists who study mood. From the beginning,
emotion researchers have been interested in categorization, developing lists of
primary emotions. Even those researchers who criticize the idea that there are
so called "basic" emotions provide lists of emotions or emotion families (e.g.,
anger, love, frustration, joy, etc.). Mood researchers, on the other hand, seem
to be less interested in the phenomenal experience of discrete moods and more
interested in reducing the mood experience to its underlying dimensions (e.g.,
pleasantness, intensity, etc.). While this difference may be the result of historical
accident, it certainly is consistent with the differences between the affective
experiences of moods and emotions. Mood is affect disconnected from its
causal object. As such, the specificity of a mood may have fewer behavioral
implications than its position on some underlying dimensions. Emotions, on
the other hand, are object oriented and the object, causal circumstances and
specific emotional reaction are therefore important for understanding and
predicting responses.
Intuition tells us that there are many different kinds of emotions, each of
which involves a unique phenomenological experience, with different
consequences for individuals and their employing organizations. As a result
of this diversity, researchers have found a need for some structure to act as
a guide for future research. The problem with emotions, however, is not a lack
of structure but, instead, a surfeit of perspectives, points-of-view, and

theoretical models. We have more structures than we could possibly use. There
are many different frameworks each based on the needs of particular
researchers (Lazarus, 1991a). Consequently, there is no one structure on which
everyone agrees (Ortony & Turner, 1990).
Plutchik (1994) has recently reviewed attempts to develop lists of "basic"
emotions. He points out that this is not a new endeavor. Descartes and Spinoza
both proposed emotional classification schemes as philopsophers attempted
to wrestle with the "passions" of humankind. All emotional positions, whether
they arise from an evolutionary perspective, a cognitive perspective, or a
psychoanalytic perspective have attempted to summarize the multitude of ways
we refer to emotions. Inherent in all of these attempts is a distinction between
primary and secondary emotions. Primary emotions refer to fundamental or
basic emotions and secondary emotions are emotional states derived from a
combination these primary emotions. This distinction is designed to summarize
the variety of emotional reactions and still retain the nuances among the many
different emotions we experience and can recognize in others.
Emotional researchers from an evolutionary perspective tend to look for
evidence of primary emotions in displays of common emotions across cultures
and also in similar displays of emotions among different species, particularly
humans and primates. This last approach is traceable to Darwin's work in 7?ze
Expression of Emotions in Man and Animals (1872[1965]).
Ekman (1992) recently applied these biological criteria to various emotion
lists and concluded that there are at least six basic emotions: anger, fear,
sadness, enjoyment disgust and surprise. Plutchik (1994) also working from
the evolutionary perspective, offers the following eight basic emotions: joy,
sadness, acceptance, disgust, fear, anger, expectation and surprise. Izard (1977),
following a more physiological approach, offers fear, anger, enjoyment,
interest, disgust, surprise, shame, contempt, distress, guilt.
Research by Shaver, Schwartz, Kirson, and O'Connor (1987) represents a
semantic classification to reduce the wide variety of experienced emotions to
a few categories (see Figure 5). This approach uses the semantic similarity of
emotion w r d s to develop categories of emotions. Shaver and his colleagues
had a group of subjects rate 213 words with respect to their "emotionness."
These ratings were then subjected to cluster analysis. Subjects identified six
categories of emotions: love, joy, surprise, anger, sadness, and fear. These six
primary categories were then meaningfully subdivided into 24 subcategories.
Eventually, all 213 words could be classified using either the 24 subcategories
or the 6 primary categories.
The Shaver et al. structure is intuitively meaningful but it is unclear what
the semantic analysis has to say about the structure of actual emotions. It may
be true that the cognitions pertaining to emotion words can be hierarchically
organized but even Shaver et al. are cautious in their willingness to extrapolate
(No subcluster
identified)
Disgust Alarm Cheerhlness Affection Disappointment

h)
C-, Envy Anxiety Contentment Longing Neglect
Exasperation Enthrallment Lust Sadness
Irritation Optimism Shame
Rage Pride Suffering
Torment Relief Sympathy
Zest
Source: Adapted from Shaver, Schwartz, Kirson, and O'Conner (1987).
Figure 5. Semantic Organization of Emotion Words

to actual emotion categories. However, one cannot help notice the degree of
overlap between the Ekman, Izard, Plutchik, and Shaver lists.
Finally, we should point out that for many researchers in the cognitive
appraisal tradition, the search for "basic" emotions is of little value. Ortony,
Clore, and Collins (1988) regard such a search as unprofitable because, in their
opinion, there is no way to choose among lists. Their approach, like other
cognitive appraisal theorists, argues for sets of relatively independent emotion
categories or families based on common appraisal processes. We will have more
to say about this appraisal process later. Here we will only point out that these
appraisal theorists still present various lists of emotions (Clore et al. talk about
such things as joy, pride, shame, distress, love, hate, etc.) as the end product
of the appraisal process..
lrnplications for Affective Events Theory
Applied psychologists are going have to await more research before the issue
of basic emotions is completely resolved. However, work to date has already
given us sufficient grounding to proceed.
First, all researchers agree that emotions can be plausibly organized into
families. These categories can provide a guiding foundation for future research.
A short list of these families would at least include anger, disgust, joy, fear,
and sadness. It might also include surprise. It should be noted that this list
is based on convergence between evolutionary and cognitive appraisal research.
These particular emotional states seem to show up time and time again-
regardless of the empirical paradigm. Consequently, at least these six emotional
families, and possibly some others, should receive future research attention.
Second, everyone concurs that specific emotional states do exist. In fact,
these states include some of our most memorable experiences, including
jubilation, loathing, grief, and terror. In every major theory, these specific
emotions are brought about by the action of cognitive appraisals. Interestingly,
appraisals are even central to theories of basic emotion. However, basic
emotion theorists generally see the appraisals as automatic, rapid, and
hardwired (Ekman, 1992). Regardless, for a person to experience an emotion
some event in the individual's environment has to be appraised. Ultimately,
appraisals involve events-in the broad sense of the word-including people
and things. Consequently, it follows that any theory of emotion must also be
a theory of how people adapt to events in their environment.
Third, some emotion terms seem to be more specific, whereas others are
more general. It is not clear whether or not the general emotions are shaped
into the specific ones. However, it is clear that some words refer to broader
states than do others. For example, we usually regret some loss that happened
in the past. However, we can be sad about some loss that did happen or that
will happen. Sadness is more general with respect to time. Likewise, fear can
refer to either avery intense or a mild emotional state, whereas the term horror
is generally reserved for something very intense. Fear is more general with
respect to intensity.
This has implications for the prediction of behavior. It seems likely that
specific emotions will prove especially useful for the prediction of specific
behaviors. Consider, for example, the emotion of sadness. How would a sad
person behave at work? That might depend on the specific variant of sadness
being experienced. A person who was disappointed over something might
behave one way. A person who felt guilty might behave another. Pity would
likely lead to the prediction of a third set of responses. According to Shaver
and his colleagues, disappointment, guilt, and pity are all part of the sadness
family. People experiencing any of these three specific emotions might all
describe themselves as sad. However, in each instance very different behaviors
are apt to result.
In spite of the difficulty of the task, organizational researchers cannot be
deterred from borrowing or developing classification schemes for describing
emotional reactions at work. If one assumes that discrete emotions have limited
and specific action tendencies, an assumption held by most emotion researchers
(see, e.g., Lazarus, 1991a), then a full understanding of the behavioral
implications of emotional reactions at work requires the use of some
classification scheme. We believe that ultimately the environment-emotion-
behavior chain will include a situation-emotion matrix that presents the key
situational features which are conducive to specific emotional reactions. It
seems to us that at the very least the analysis of the structure of work emotions
is as important as the analysis of the dimensions of job satisfaction.
The Structure of Mood
So far we have been discussing mood as if it were a single thing. However,

experience tells us otherwise. Intuitively, when we describe our mood we tend
to do so in terms of two words: "good" and "bad." In psychological jargon,
we might say that people articulate their feelings with reference to hedonic tone,
"positive" for "good" and "negative" for "bad." We also categorize our moods
in terms of their intensity. We can say, for example, that we feel "okay" or
"very good." Both are generally positive, but the latter is more so.
Researchers, of course, attempt to describe mood more systematically and,
as we have said, mood researchers have been particularly concerned with
reducing the mood experience to underlying dimensions. The most common
paradigm for doing this task has been to administer a large number of mood
items to a group of respondents. These responses are then subjected to a
dimensional analysis (cluster analysis, factor analysis, etc.) (Mayer & Gaschke,
1988; Watson & Tellegen, 1985). The obtained factors/clusters are taken to
indicate the underlying dimensional structure of mood. Although this and
similar paradigms are widely used, the dimensions obtained vary somewhat
among researchers. In general, two different two-dimensional structures have
received wide currency. One structure, frequently adopted in the organizational
sciences, conceptualizes mood in terms of Positive Affectivity (PA) and
Negative Affectivity (NA). Another widely discussed structure, less frequently
used in organizational research, dimensionalizes mood states in terms of
Hedonic Tone (with positive and negative affect as anchors on a single
continuum) and Intensity. The evidence for each of these models is reviewed
below.
Positive and Negative Affectivity
The empirical literature on state and trait mood has been extensively
reviewed by Watson and Clark (1984) and Watson and Tellegen (1985).
According to these authors, both state and trait mood can be best represented
in terms of two distinct dimensions: Positive Affectivity (basically good
feelings) and Negative Affectivity (basically bad feelings). At the high pole,
PA is characterized by such adjectives as "energetic," "exhilaration," and "joy."
People who report high positive affectivity exhibit a zest for living. At the low
end, however, low PA is not the presence of negative affect. Rather, the low
pole is characterized by the absence of positive affect. Individuals scoring low
on PA are best seen as apathetic and listless. NA manifests itself in a different
complex of feeling states. People score high on NA when they report anger,
nervousness, anxiety, guilt, sorrow, and so on. At the low pole, NA does not
involve the presence of positive affect. Rather, it involves the absence of
negative affect. In this sense, it functions in a manner analogous to PA.
Individuals scoring low on NA often report being placid and content.
Seminal work on this NA/ PA model of mood was conducted by Zevon and
Tellegen (1982). These researchers had 23 subjects fill out the same mood
adjective checklist for 60 consecutive days. Afterwards, they conducted 23
within subjects factor analyses. For 2 1 of the 23 individuals, two strong and
independent factors emerged, one characterized by positive affect and the other
by negative affect. These data provided clear support for the P A / N A model.
Since the initial work of Zevon and Tellegen (1982), several others studies
have uncovered the two basic PA and NA dimensions (e.g., Gotlib & Meyer,
1986; Watson & Clark, 1991; Watson, Clark, & Carey, 1988; Watson, Clark,
& Tellegen, 1988; Watson & Tellegen, 1985). For example, one representative
study was reported by Watson (1988). In this study 80 undergraduate students
were administered a daily mood checklist for 6-8 weeks. These mood reports
were than factor analyzed and the PA and NA dimensions were again
recovered. Watson (1988) also reported that social activity and exercise were
most strongly related to PA, while perceived stress was strongly related t o NA.
Physical symptom reports were related t o both NA and PA.
Affectizw Events Theory 25
Although still limited, cross-cultural investigations have also produced

consistent findings. For example, using a method similar to the one reported
above, Watson, Clark, and Tellegen (1984) uncovered the same two factor
structures in a sample of Japanese citizens. Similar results were also obtained
in another cross-cultural study by Almagor and Ben-Porath (1989).
The NA/ PA model of mood structure has been extremely influential within
the organizational sciences (see George, 1992, for a review). Nevertheless, it
is not the only useful way for dimensionalizing mood states. A second model
has also obtained a great deal of empirical support. It is to this alternative
approach that we now turn.
Hedonic Tone and Intensity
As with NA/PA model, various researchers have argued that mood can be
thought of as involving two dimensions but have proposed a different set of
dimensions (for reviews see Larsen & Diener, 1992; Russell, 1979). The first
dimension of this alternative structure is sometimes referred to as Hedonic Tone
and other times as Pleasantness. The Hedonic Tone or Pleasantness dimension
describes mood states as falling on a scale from very positive to very negative
with many points in between. The positive pole of this Hedonic Tone dimension
would include such things as "happy" and "carefree," while the low pole would
be characterized by such markers as "sober" and "distressed." Note that this
is a single bi-polar continuum.
The second dimension in this alternative structure is the level of intensity.
That is, feelings can range from very intense to very mild. So, for example,
we can see the high end as including such adjectives as "restless" and
"changeable," while the low end is characterized by such adjectives as
"controlled" and "peaceful."
As with NA and PA, the Hedonic Tone and Intensity structure has received
empirical support. Much of this support comes from the work of Russell and
his colleagues who have found evidence for two broad dimensions that he calls
"Degree of Arousal" and "Pleasantness-Unpleasantness" (e.g., Russell, 1978,
1979; Russell & Mehrabian, 1977; Russell & Ridgeway, 1983).
W h y the Confusion Over Factor Structures?
At first glance the two previous sections appear to completely contradict

one another but, in fact, it is possible that both models can provide a good
mathematical fit to the data. To understand why this is so one needs to consider
that the NA/ PA and Hedonic Tone/ Intensity solutions have in common the
fact that each describes mood in terms of two orthogonal factors. Both models
provide a simplified representation of a more complex reality. Let us look at
this in more detail.
's'
v,
Several authors have noted that when words referring to mood are judged
in terms of similarity, the relationships among the words can be visually
depicted in terms of a circular or "circumplex" model (see Figure 6) (Larsen
& Diener, 1992; Mayer & Gaschke, 1988; Meyer & Shack, 1989; Russell, 1980;
Watson & Tellegen, 1985).
What is especially important for our purposes is how one might represent
this circumplex in two-dimensional space. Given the circular relationship
among the words, multiple two-dimensional structures are possible. As shown
in Figure 6, the two structures we have been discussing are simply different
axis rotations of the same data. Both summarize the relationships equally well
in that both structures can be used to precisely locate a given mood in the
dimensional space.
Because, mathematically speaking, both models generally fit the observed
data the value of either approach cannot be decided on mathematical grounds
alone. Rather, the worth of either solution becomes, not so much a matter
of truth, as a matter of methodological and theoretical clarity. Consequently,
both "camps" have looked to other evidence to make the case for their preferred
structures. We will review some of these arguments.
Robustness Across Multiple Operations
As we have seen, both the NA/PA and the Hedonic Tone/Intensity

dimensions have been recovered from factor analytic studies, of survey
responses. However, the Hedonic Tone/Intensity model has been
demonstrated with other paradigms as well. For example, much of the work
of Russell and his colleagues has used multi-dimensional scaling techniques
(e.g., Russell, 1979; Russell & Ridgeway, 1983). Similarly, the Hedonic Tone/
Intensity solution also shows up when pictures and not words are used as
stimulus materials (Russell, Lewicka, & Niit, 1989). Conversely, evidence for
the NA/ PA model has been limited to factor analytic research on verbal scales.
Conceptual Descriptions
There has been some conceptual ambiguity regarding the two dimensions
of NA and PA that is relevant to choosing an appropriate dimensional
structure. Essentially, most adjectives that load on either NA or PA describe
only the high pole for each adjective. Take, for example, the case of the Positive
and Negative Affectivity Schedule (or PANAS, see Watson, Clark, & Tellegen,
1988). The PANAS is probably the most widely used measure of PA and NA.
The schedule was constructed by using exploratory factor analysis to identify
items with high loadings on one dimension (either NA or PA) and low loadings
on the other. Interestingly, this procedure left the scale's authors with only items
indicative of high positive affect (such as "energetic") or high negative affect

(such as "anxious").
Other factor analytic research has also produced mixed results. On the
supportive side, Mayer and Gaschke (1988) report that "drowsy" and "tired"
had a highly negative (-.39 and -.36, respectively) loading on PA and very
low loadings on NA (-.04 and -.09, respectively). Similarly, "calm" was
adequately loading on NA (-.59) and had a low loading on PA (.06). The results
for "content" were less definitive but still supportive. "Content" had a lower
loading on PA (.39) than NA (-.43). In the Mayer and Gaschke (1988) study,
therefore, there were at least a few clean markers of low PA and low NA.
However, not all of the research has been this supportive. In one study,
Meyer and Shack (1989) collected self-ratings on a variety of different mood
adjectives. These data were then factor analyzed and subjected to a varimax
rotation. Theoretically speaking, several adjectives would seem to have been
good markers for low NA, including such things as "relaxed," "calm,"
"pleased," "content," "satisfied," "quiet," and "still." However, none of these
items produced clear results. Most contained similar loadings on both NA and
PA. In fact, "pleased," "content," "satisfied," "quiet" all had higher absolute
loadings on PA than they did on NA. Of the remaining three that had higher
loadings on NA, the clearest solution was for "calm," which loaded -.27 on
NA and .19 on PA. Thus, "calm" was the only reasonable marker of low NA
while it was also more or less independent of PA. Of course, -.27 is not a
decisive loading.
Similar results were obtained for Meyer and Shack's low PA adjectives. Four
items seemed to be good candidates for low PA: "sluggish," "drowsy," "sleepy,"
and "quiet." However, the first three of these adjectives all had higher (positive)
absolute loadings on NA. That is to say, they were better markers of high
negative affect than they were of low positive affect. In fact, the item that came
closest to being a clean marker of low PA was "quiet." It had a -.31 loading
on PA and a .26 loading on NA. Every other item with a negative loading
on PA had a higher (and positive) loading on the NA dimension. Put
differently, there were no clear indices of low PA.
NA and PA may be more independent at the high pole than at the low pole.
This phenomenon has long been known. For example, Zevon and Tellegen
(1982, 'p. 112) maintained that NA and PA were "descriptively bipolar but
affectively unipolar dimensions." To state the matter differently, NA and PA
seem oniy to be independent constructs at the high end of their poles.
What this means theoretically has yet be fully articulated. It is difficult to
understand a dimension that is one thing on the low end and two things on
the high. Certainly, this makes the NA/PA model more difficult to interpret,
as most constructs have two poles, not one. The Hedonic Tone/ Intensity model
avoids these concerns. Hedonic Tone and Affect Intensity are (at least in
principle) orthogonal at both the high and low poles. Thus, understanding this
structure is more straightforward.
The Issue of Error Variance
A series of studies by Green, Goldman, and Salovey (1993) is also

problematic for the N A / P A model of mood. These authors conducted
confirmatory factor analyses of various mood adjectives. They found that when
measurement error was taken into account, the two factor N A / P A model
provided a poor fit to the data. A better fitting model involved a single bi-
polar dimension of mood, with P A and NA fitting as opposite ends of this
continuum. This bi-polar dimension, of course, is similar to the factor of
Hedonic Tone. Similar arguments have been made by Russell (1979).
The Issue of Time Frame
As noted above, it has been argued that NA and P A are basic dimensions
of both state and trait mood (Watson, 1988). However, some research suggests
that the structure of mood may vary based upon the time perspective taken
by the respondent. For example, studies by Diener and Emmons (1984) and
Diener and Iran-Nejad (1986) were able to replicate the independent NA and
PA dimensions only when individuals were asked to describe their mood in
general. When subjects were asked to describe their current mood state (as
opposed to their general predispositions), negative and positive affect exhibited
a significant, negative correlation. Diener and his colleagues argued that it is
difficult to feel both good and bad at the exact same time. If this is so, then
NA and P A might be useful personality dimensions, but less useful as
dimensions of state mood.
It should be emphasized that not all of the data are consistent. For example,
Watson (1988) and Mayer and Gaschke (1988) were able t o recover the NA/
PA factor structure with judgments of one's present mood. The point here is
that the NA/ PA structure is only inconsistently obtained with ratings of current
mood, while the Hedonic TonelIntensity structure seems to be more stable.
It would seem that the more stable factor structure is likely to be the more
useful one.
Other evidence seems to suggest that people tend not to experience both
types of affect simultaneously. Experimental and field evidence presented by
Baron (1976, 1984) suggests that when incompatible mood states exist the
stronger one tends to cancel the weaker. Baron (1976, 1984), in fact, has found
that these incompatible responses are a useful tool for conflict management.
Once again, we should emphasize that independent mood states d o not
necessarily imply independent mood traits. Personality dispositions to
experience moods may well be interdependent. This is a separate empirical
question. Our concern here is with mood states. Both experimental (Baron,
1984) and correlational (Baron, 1976; Diener & Emmons, 1984; Diener & Iran-
Nejad, 1986) evidence suggests that individuals experiencing a positive mood
state will not be experiencing a negative one. Thus, when mood is assessed
as a state, there is at least some evidence that NA and PA are (negatively)
correlated. This would suggest the greater usefulness of an Hedonic Tone/
Intensity Structure over the NA/ PA structure.
integration .With Other Research
Watson and Clark (1984), recognizing the inability to choose between the
two structures on mathematical grounds, argued that a more informed choice
could be made by looking at the way in which the two structures fit with other
relevant literature. The NA/ PA model clearly has some strengths in this regard.
As noted by Watson and Clark (1984), Watson (1988) and Watson, Clark,
and Tellegen (1988), the same NA/ PA factor structure can be used to describe
both state and trait mood. NA corresponds to the personality dimension of
Neuroticism and PA to the dimension of Extraversion. Thus, the use of the
NA/ PA solution allows for a clear connection between states and traits. On
the other hand, as Larsen and Diener (1987) have shown, very similar links
can also be established between Hedonic Tone/ Intensity and other personality
dimensions such as Affect Intensity. Similarly, the Hedonic TonelIntensity
model is also consistent with much research on attitudes (Osgood, Suci, &
Tannenbaum, 1957). Consequently, the advantages of either structure using
this criterion are unclear.
For purposes of Affective Events Theory, there is at least one respect in which
the Hedonic TonelIntensity model might be more useful. It is our belief that
a common structure for moods and emotions would be advantageous and that
the Hedonic Tone/ Intensity structure better fits that bill. As we will describe
in the next section, emotional states result from a two stage appraisal process
(Frijda, 1986; Lazarus, 1991a, 1991b). The first part of the appraisal involves
an assessment of the "goodness" or "badness" of the event, as well as an
assessment of the event's importance. Appraisals of "goodness" lead to positive
affective states, the intensity of which is relatively high for important qvents
and relatively low for unimportant events. Similarly, appraisals of "badness"
lead to negative affective states, whose intensity is likewise modulated by
importance. We shall have more to say about emotion formation later. For
now it should suffice to indicate how closely the appraisal of "good-bad"
corresponds to Hedonic Tone, while the appraisal of "important-unimportant"
impacts intensity. The observations suggest that the Hedonic Tone/ Intensity
model might be more parsimonious and more easily integrated with various
theories of emotion.
Affectizle Ez)eltts Theory 31
THE EFFECTS OF EVENTS ON EMOTION AND MOOD

Emotion Generating Events
Our theory gives primary emphasis to the role of events as proximal causes
of affective reactions and then as more distal causes of behaviors and attitudes
through affective mediation. Consequently, at this point we feel obligated to
discuss what we mean by event. This obligation is easily met because we mean
nothing more than what the word event means in ordinary language. The
World Book Dictionary gives the primary definition of event as "a happening,
especially an important happening." The Random House Dictionary goes on
to add "something that occurs in a certain place during a particular period
of time." Implied in both of these definitions is the idea of change, a change
in circumstances, a change in what one is currently experiencing.
Some, but by no means all, events have affective significance in that they
generate an emotional reaction or mood change in people. It is these changes
that we need to focus on. T o d o so requires that we try to answer at least two
questions. First, what kinds of changes have affective significance? Second.
how do specific representations of events eventuate in the experience of specific
emotions?
It turns out that these two questions correspond nicely to a two stage
appraisal process advocated by most cognitively oriented emotion researchers
(Lazarus, 199 la, 199 1b; Ortony, Clore, & Collins, 1988; Roseman, 1984; Smith
& Ellsworth, 1987; Stein, Tribasso, & Liwag, 1993). We present an overview
of the common elements of their theories to describe the processes involved
when events elicit emotional reactions. We should say from the outset that
emotion theories come in many different forms. Plutchik (1994) has described
four different categories: motivational theories, psychoanalytic theories,
evolutionary theories, and the cognitive theories which we focus on. All theories
assume that emotional reactions generally begin with an appraisal of an event
(Plutchik, 1994); however, cognitive theories focus their attention on just this
point. Because this is the issue we need to focus upon, we feel that the cognitive
theories have more to say t o us. However, we invite our readers to examine
other mediators of event-emotion relationships.
There appears to be a common emotion elicitation process at the core of all
cognitive appraisal theories. This process begins with an event which is initially
evaluated for relevance to well being in simple positive or negative terms. This
initial evaluation also contains an importance evaluation which influences the
intensity of the emotional reaction. Initial appraisal leads to further, more
specific, appraisal of context focusing on consequences, attributions, coping
potential, and so forth. While different theorists offer different appraisal
dimensions, all suggest that it is this secondary level of appraisal which results
in the experience of discrete emotions like anger, sadness, or joy.
Prima y Appraisal
Initial or primary appraisal appears based on "concern relevance" (Frijda,

1993), relevance to well being. Just what does this mean? Most theorists suggest
that concern relevance is intricately tied to one's personal set of goals and
values. Lazarus (1991a) is very specific when he says that "harms or benefits
depend on goal commitments which reside in the person and are either thwarted
or facilitated by the behavior of the environment" (p. 92). He concludes that
initial appraisal involves an assessment of "goal relevance," does the event touch
on some issue of personal desire or concern, and "goal congruence," is the event
consistent (beneficial) or inconsistent (harmful) with those desires or concerns.
Similar positions are offered by Ortony, Clore, and Collins (1988), Frijda
(1993), Stein, Tribasso, and Liwag (1993), among others. Recently, Berkowitz
(1989) has reformulated the classic frustration-aggression theory of Dollard,
Doob, Miller, Mowrer, and Sears (1939) in these terms, concluding that
frustration is a blockage of goal gratification and the aggression effect is i
9
mediated by negative affective responses. z
In addition, theorists who argue that goal relevance is essential to the 4
f
emotional reactions to events generally add that the intensity of the emotion
is directly correlated with the importance or desirability of the goal. Ortony,
i
Clore, and Collins (1988) add other variables which influence the intensity of
the emotional reaction to an event, including unexpectedness and existing
arousal levels.
Although both positive and negative goal relevant events can occur,
producing positive and negative emotional reactions, respectively, Taylor
(199 1) reviews evidence which suggests that the effects of positive and negative
events are not symmetrical. More specifically, she concludes that negative
events produce stronger reactions than do positive events. Although the
research which examines this issue suffers from the difficulty of equating event
strength, the body of the work suggests that negative events produce stronger
physiological responses and stronger subjective feelings of affect.
The connection between the goal relevance involved in general emotional
appraisals and the motivational and affective consequences of work goals is
obvious. Locke and Latham (1990) have provided a useful discussion of the
relationship between work goals and affective reactions and we will add only
a few points relevant to emotional appraisal. First, the type of goals relevant
to emotional appraisal go well beyond the performance goals which form the
core of organizational study in this area. People have a wide variety of goals
and objectives. These preferred states can be what people strive for, what they
seek to avoid, what they hope to maintain, what they want to see occur, and
so forth (Cropanzano, James, & Konovsky, 1993). All are relevant to emotional
appraisal.
Affectizie Events Theory 33
Second, in agreement with most cognitive appraisal theorists (Lazarus,

1991a; Ortony, Clore, & Collins, 1988; Stein, Trabasso, & Liwag, 1993) we
assume that goals are hierarchically organized (cf. Cropanzano et al., 1993).
People have broad, distal goals as well as proximal subgoals which are
instrumental to the attainment of these higher level objectives. In addition, the
goal-subgoal connections within hierarchy structures can be described with
different types of instrumentality. Some subgoals are necessary but not
sufficient for the attainment of broader objectives, as, for example, the
instrumentality of a college degree for professional success. Some subgoals are
sufficient but not necessary for the attainment of broader objectives, as, for
example, winning the lottery as a way of achieving wealth. We expect that
the importance of any particular goal, and therefore the intensity of emotional
arousal, is influenced by the position of the event implicated goal in the
hie~archyand well as the nature of the instrumental relationship with other
goals. For example, it might be that the failure to obtain a necessary but not
sufficient goal may have greater negative implications than the actual
achievement of that goal has positive emotional implications. Similarly,
meeting a sufficient but necessary goal may have greater positive emotional
significance than the failure to meet that goal would have negative significance.
Third, people can and do focus on different elements of their goal structures
at any particular time. It is logical to assume, therefore, that goal attention
influences the reaction to specific events.
We wish to make one last point on this issue of goals and emotional
appraisal. After reading what we have written about the nature of emotional
appraisal, one might argue that the appraisal of the goal significance of events
greatly resembles the process of standard comparison which we ascribed to
the Cognitive Judgment approach. The similarity is not lost on us, but we must
point to some important differences. To begin with, our focus here is on the
formation of discrete emotional reactions not general evaluations. Second, our
focus is on the appraisal of the emotional significance of an event, not the goal
significance of a feature of the environment. Third, with regard to emotional
experiences, goal significance is just the beginning of the full appraisal process.
This all being said, one fundamental point of similarity remains and that is
the ubiquity of goals as frame of reference for evaluating one's personal state
of affairs.
Secondary Appraisal
All cognitive appraisal theorists argue that primary appraisal is followed by

a secondary appraisal, an interpretive "meaning analysis" (Smith & Pope, 1992)
in which specific cues from the environment and the person are evaluated and
discrete emotional responses elicited. where the theories differ is in the specific
dimensions they propose as relevant to the appraisal process. So, for example,
Smith and Ellsworth (1985) proposed that emotional events are appraised in
terms of five dimensions: pleasantness, certainty, self vs. other agency,
attentional direction (toward or away from), and effort needed to deal with
the event. Roseman (1984) proposed that events are appraised in terms of
whether the situational state is consistent or inconsistent with motives, certainty
of outcome, agency, motivational state (consistent with a positive or negative
motive), and coping potential. Lazarus (1991a) offered blame or credit (who
is responsible), coping potential and future expectancy of the situation
improving or getting worse as the key appraisal dimensions. Other dimensional
structures have been offered by Frijda (1987), Ortony, Clore, and Collins
(1988), and Scherer (1984) among others.
Although most appraisal theories do not specify any particular sequence to
secondary appraisal, essential elements of the process can be captured by using
a decision tree as a model. Figure 7 illustrates this using the configuration set
forth by Stein, Trabasso, and Liwag (1993). We use this configuration as an
example, without endorsing this dimensional structure over any other.
Research on the validity of this appraisal process can be illustrated by a
recent study by Roseman, Spindel, and Jose (1990). They asked subjects to
recall events for 16 specific emotions and then to respond to a set of appraisal
questions for each. They showed that the responses to the appraisal questions
were able to differentiate among the specific emotions. Other studies done by
Weiner (1985), Smith and Ellsworth (1985), and Frijda (1989), among others,
have generally supported each researcher's conception of appraisal dimensions,
although no definitive conclusion about the usefulness of one configuration
over another can yet be made.
Mood Generating Events
There exists quite a substantial empirical literature which speaks to the

antecedents of moods. Interestingly, this literature exists primarily as a
catalogue of mood manipulations in studies conducted to examine some
hypothesis about mood effects. Most of these manipulations seem rather
intuitive and ad hoc, allowing for little to be gleaned about their underlying
conceptual similarities. Theoretical discussions of the antecedents of mood
have been far less frequent than have similar discussions about the antecedents
of emotions. All this is meant to indicate that there is much less to say about
the antecedents of mood then there was to say about the antecedents of
emotion.
Nonetheless, Morris (1989) has contributed a useful discussion of this topic
which we will summarize. He suggests that there are four positions on the
sources of experienced moods. Position 1 states that moods are the result of
mildly positive or mildly negative events. This is the position which guides most
of the laboratory research on the effects of moods. Indeed, experimental studies
Precipitating
Event - Initial State
I
Attend To Event
I
Identify and Understand Event
I
Assess Relevence Of Event To Valued Goals
/ ---l
Obstructs Goals
Facilitates Attainment I
of Goals Assess Certainty of
I Goal Failure
I
Assess ~ektaintyof
Determine If Plan Is Available
Goal Success
I I
Certain No Plan
I 1
Source: Adapted from Stein, Trabassb. and Liwag (1993).
Figure 7. Appraisal Flow for Happiness and Sadness

have demonstrated the efficacy of manipulating mood by showing brief videos,

providing subjects with cookies, playing pleasant or unpleasant music, helping
subjects "find" a small amount of money, inducing success on experimental
tasks, and so forth. While these studies show that mild events of "hedonic
relevance" induce a mild affective state more consistent with moods than
emotions they do little to tell us what it is about these events that produce
the state. Morris says they are events of hedonic relevance, but this says nothing
more than they induce affect. We acknowledge their effectiveness but also
suggest that however intuitive they may seem a fuller statement of what kinds
of events are of mild hedonic relevance is absent.
The second position noted by Morris is that moods result from the offset
of emotional reactions. Whether they are believed to be residues of the
emotions, simultaneous affective experiences initially overwhelmed by the
emotion or opponent process reactions to the end of the emotion varies. In
each case, the mood follows the emotion. Morris (1989) says that it is difficult
to comment on the validity of this position because there are too few relevant
empirical studies.
Position 3 states that moods can result from the recollection of emotional
events. It seems clear that mood states can be directly affected by cognitive
processes of recall and imagination in the absence of actual, current affective
events. Many popular mood manipulations count on it. S o too d o authors
and screenwriters. Morris (1989) argues that the affective response is likely to
be "mood like" in that it will be of low intensity with no specific response
tendencies. While we d o not argue that moods can result from recall or
imagination we would add that so too can full blown emotional responses.
Many people know all too well that thinking about a deceased loved one can
generate real sadness, not undifferentiated background affect.
Finally, Position-4 states that inhibition of a full blown emotional response
can result in a residual mood. Wharton and Erickson (1993) argue that
"emotional display rules" exist in organizations and that there are increasing
expectations t o control emotional displays at work. T o the extent that Position
4 is valid, these expectations of emotional suppression may increase the
prevalence of negative moods, with accompanying consequences.
OTHER CAUSES OF AFFECT LEVELS AT WORK

Dispositional Influences on Affect
Thus far we have presented evidence that affective traits are one cause of
job satisfaction. We have further maintained that these traits are partially
rooted in the individual's biology, and may have some genetic basis as well.
Despite this, we were careful to note that these observations do nothing to

rule out the important role played by situational determinants.
Main Effects of Affective Dispositions
Research suggests that affective dispositions exert main effects on job

satisfaction. One of the best illustrations of this is a large survey of 1,8 16 drivers
conducted by Czajka (1990). Czajka (1990) reports that job satisfaction was
predicted by both positive affectivity (r = 0.43) and negative affectivity (r =
-0.46). Additionally, after controlling for salary and tenure, NA and PA
together account for 29% of the variance in job satisfaction. Czajka's (1990)
findings are important for our present purposes, as he reports no evidence of
a disposition by situation interaction. Instead, situational variables (i.e., pay
and tenure) and person variable (i.e., NA and PA) each contributed additive
main effects. Similar results were obtained in two studies by Cropanzano,
James, and Konovsky (1993). Findings. of this type would suggest that both
affective traits and situational attributes impact job satisfaction. However, a
closer look suggests that reality might be a little more complicated than this.
Affective traits appear to act as latent predispositions that help set the stage
for individuals to have more or less intense bouts of emotion. These traits are
affective predispositions and not the experience of affect, per se. Thus, we can
further see that a given affective trait manifests itself only under particular
environmental conditions. This is to say, for example, that an individual high
in trait Negative Affectivity or neuroticism, need not go through life with a
chronic sense of discontentment. Rather, such an individual is predisposed to
react more strongly to negative events when they happen to occur. When no
negative event takes place, individuals high and low in trait NA should have
similar levels of mood and job satisfaction.
This observation has some interesting conceptual implications. If affective
dispositions are defined as manifestations of repeated bouts with negative and
positive emotion, and if such bouts of emotion are partially dependent upon
the situation, then affective traits are (partially) defined in reference to the
environmental context. Concrete evidence for this can be gleaned from a study
by Werner and Pervin (1986). Werner and Pervin content analyzed the items
from six widely used personality inventories. This included, but was not limited
to, scales that measured dispositional affectivity. Across the six inventories,
55.7% of the items made specific reference to situations. Thus the personality
traits were being measured, and to some extent defined, as to how they were
manifested in particular settings.
This does not mean that researchers can or should not study traits and
environments separately. In fact, the very complexity of the problem demands
that we do so. Nevertheless, a full understanding of the way in which affective
dispositions manifest themselves demands that we simultaneously consider
both the trait and the eliciting situation. Below, we consider this issue from
two perspectives. First, we consider statistical interactions between affective
dispositional and situational stimuli. Second, we explore the reciprocal
transactions between the person and the environment.
Stntisticnl lnteractiorls
Bolger (1990) surveyed pre-medical students in the period leading 1.p to and
shortly after taking the MCAT exam. Bolger also assessed each individual's
level of neuroticism. Neuroticism is closely related to trait NA (Meyer & Shack,
1989; Watson & Clark, 1984). Bolger (1990) found that individuals high in
neuroticism were more reactive to the stress of the exam. In particular, the
high neuroticism individuals reported more anxiety in the week preceding the
MCAT, but were not more anxiety ridden in other weeks. Negative emotion
was only manifested as the stressor approached. When the stressor was further
away in time, all individuals were about equally low in mood negativity.
Similar findings were also obtained in another longitudinal study by Bolger
and Schilling (1991). Within a community sample of adults, Bolger and
Schilling (1991) found that participants high in neuroticism had much more
negative responses to various daily hassles, such as interpersonal conflict.
In a field study of working individuals, Parkes (1990) drew similar
conclusions. When placed in a stressful environment, teachers who were high
in NA reported more symptoms of distress than did teachers low in NA.
Conversely when the environment was less stressful, teachers reported similar
levels of distress, regardless of their level of negative affectivity.
All of these findings were replicated and extended by Marco and Suls (1993).
These later researchers also discovered that individuals high in trait NA were
more reactive to negative events. Moreover, Marco and Suls (1993) also found
that high NAs take longer to recover from a stressor. Once again, the emotional
outcomes associated with NA were only present when a negative event
occurred. At other times, individuals high in trait NA demonstrated levels of
emotion that were comparable to their low NA counterparts.
These ideas were experimentally tested by Larsen and Ketelaar (1991).
Larsen and Ketelaar (199 1) first assessed undergraduate students on their levels
of neuroticism and extroversion. (PA is highly correlated with extroversion,
but extroversion includes other things as well. For a discussion, see Meyer and
Shack [1989] and Johnson'and Ostendorf [1993]). Larsen and Ketelaar (1991)
then exposed subjects to three kinds of stimuli: positive affect provoking,
negative affect provoking, and nonaffect provoking (neutral). As one might
expect, individuals high in neuroticism reacted primarily to the negative events.
They were relatively nonreactive to the neutral and positive stimuli. Conversely,
individuals high in extroversion reacted mainly to the positive events. They,
in turn, were relatively nonreactive to the negative and neutral stimuli. Again,
we see evidence that affective traits serve as predispositions to respond within

a particular environmental context.
It is noteworthy that this reasoning has been directly applied to job
satisfaction. In an experimental study Bittle and Hausenstein (1990) found that
dispositional affectivity was only related to job satisfaction when the work
environment was generally unenriched and negative. However, while in a
relatively enriched environment individuals were generally satisfied with their
work, regardless of whether or not they were predisposed to negative affectivity.
Trait affectivity does seem to be related to both one's mood and also one's
level of job satisfaction. However, in both cases, the personality trait acts as
an affective predisposition. That is, it predisposes people to respond with
greater or lesser intensity to either a positive or a negative event.
The Dynamic Transaction Between the Person and the Environment
It is important to note that individuals are not passive recipients of

environmental pressure. Instead, individuals move through their lives both
influencing and being influenced by their environments. This is true for all
personality traits, but we are here concerned with the evidence pertaining to
affective predispositions. Unfortunately, virtually all of the research has been
conducted with only predispositions for negative emotion. It is likely (albeit
I still unclear) that trait PA behaves in a similar fashion. In any case, these
limitations in the literature have caused us to limit our discussion to trait NA
and related dispositions.
Individuals high in trait negative affectivity behave differently than their
counterparts who are low on this trait. In particular, they are more likely to
engage in contentious interpersonal tactics by being obstinate or
argumentative. For example, Buss, Gomes, Higgins, and Lauterbach (1987)
found that individuals predisposed to negative emotion were more likely to
quarrel with their spouses.
1 These findings were extended in a longitudinal field study by Bolger and
Schilling (1991). In various settings, including the workplace, individuals high
in neuroticism were more likely to argue and quarrel with others. Bolger and
Schilling (1991) also found that this tendency to fight was actually one cause
1 of the experienced negative mood. That is, by their contentiousness, individuals
high in neuroticism elicited hostility from others. This hostility, in turn, caused
them to experience negative emotion.
Although little direct evidence exists, it also seems possible that the work
environment can affect individual levels of trait NA. In one longitudinal study,
? Kohn and Schooler (1982) found that trait levels of personal distress were
accentuated by oppressive working conditions. For example, workers forced
to labor under heavy work loads with tight deadlines were likely to show higher
levels of trait-based distress. Kohn and Schooler's (1982) study is highly

suggestive. However, more research is needed to replicate and extend these
findings.
Environmental Causes
Environmental psychologists have uncovered a variety of factors which can

change the level of reported affect (see Bell, Fisher, & Loomis, 1978, for a good
review). Such environmental conditions as weather, air pollution, noise, and
negative ion level appear to influence affect and related behaviors. Generally
these factors operate in the background, influencing mood states rather than
specific emotions.
In general, several aspects of weather are related to self-reported mood.
Laboratory studies have shown that exposing subjects to uncomfortable levels
of heat produces negative mood states (Bell, Garnand, & Heath, 1984; Griffitt
& Veitch, 197 1). Similarly, Cunningham (1979) found that a hot temperature
during the summer reduced mood levels while in the winter, a warmer
temperature was associated: with a better mood. It has been shown that sunshine
has a positive effect on mood states (Persinger, 1975) and high humidity is
associated with negative affect (Cunningham, 1979; Persinger, 1975).
Laboratory research has shown that exposing people to noxious pollutants
can worsen their moods (Rotton, 1983) but the picture is complicated by
research indicating that different pollutants have different effects during
different seasons. For example, Cunningham (1979) found that higher levels
of carbon monoxide were associated with better moods during the summer
and worse moods in the winter. Obviously, these results and other results like
them must be interpreted in the context of severe potential for spurious effects
of other variables. Baron, Russell and Arms (1'985) have shown that negative
ions in the environment affect mood levels. However, in their research the
direction of the effect depended upon other environmental stimuli.
Crowding is another environmental variable that can have negative
consequences for mood levels (Freedman, 1975). Oldham and Fried (1987)
found that perceptions of crowding were associated with lower levels of job
satisfaction. However, the negative mood effects of crowding can be reduced
when people believe they have personal control (Fleming, Baum, & Weiss,
1987).
Overall, the evidence suggests that a wide variety of environmental factors
influence individual affect levels. By and large, these operate in the background
to influence mood levels but it seems clear that their consequences on
organizational behaviors, as mediated by mood states, are likely to be
important.
EMOTION EPISODES:
THE EBB AND FLOW OF EMOTIONAL EXPERIENCE
Our own discussion of emotions, guided as it is by theories and research on
the topic, has so far characterized emotions as discrete reactions precipitated
by specific events. Although we believe this to be a useful characterization of
any specific emotional reaction, in a very important sense it misses a
fundamental aspect of emotional experience. Frijda (1993) notes that when
people are asked to describe an emotional experience, often they d o not simply
\
report a single emotion precipitated by a single event. Instead they report a
series of emotional transactions with the environment, all coherently organized
around a single underlying theme. Frijda refers to this coherent and dynamic
series of emotional experiences as an emotion episode, a situation in which
a single event of affective significance leads to the unfolding of a series of
subevents, also with affective significance. Each of the subevents can produce
a distinct, even opposite, emotional response but the full episode is driven by
what Lazarus (1991a) would call a core relational theme.
Frijda (1993) goes on t o say that during this episode the person remains
in a state of "continuous emotional engagement" (p. 387). Emotional
engagement refers to a heightened level of arousal and attention. T o us, this
suggests that during the emotion episode, the person's attention is focused on
issues related t o the underlying theme, possibly leaving fewer resources to
commit to job performance. It also suggests that small events, coworker
comments, organizational memos, and so forth take on increased, and perhaps
unwarranted, emotional significance. Finally, it suggests that people may
overreact to emotional events unrelated to the underlying core theme as their
heightened level of arousal produces a misrepresentation of the event's
emotional import. A likely consequence of emotional engagement is that people
will experience more intense and diverse affective reactions then they would
otherwise.
The effects of emotional engagement on affect levels during emotion episodes
can be seen in research examining Zillman's Excitation Transfer Theory
(Zillman, 1979). Excitation Transfer Theory proposes that individuals'
emotional experiences are enhanced when people are already in aroused states,
regardless of the source of the original arousal or of its initial hedonic direction.
However, this excitation transfer should only occur when the arousal
experience has become "disconnected" from its original cognitive label.
Supporting data is found in a number of studies by Zillman and his associates
(e.g., Cantor, Zillman, & Bryant, 1975; Zillman & Bryant, 1974; but also see,
Branscombe, 1985). Of course, the persistent effects of mood on how we
Interpret the world (Morris, 1989), which we will discuss later, also illustrates
the affective consequences of continuous emotional engagement.
lmagine a worker who hears about the possibility of a large scale layoff in
his organization. This event, being of relevance to his well being, is likely to
elicit an emotional reaction. Because it has negative implications it will elicit
a negative reaction and because it involves the anticipation of future harm that
negative reaction is likely to be fear or anxiety. This layoff possibility, in turn,
instigates a series of subevents, many caused by the coping mechanisms of the
worker himself, each of which can have some emotional significance. For
example, the worker may talk to his supervisor who tries to reassure him of
his worth to the company and these reassurances produce a positive affective
state. The worker may talk to other workers who relate tales of downsizii~g
or outplacement, each with emotional implications. During the "episode" the
worker will experience wide swings of affect in both positive and negative
directions. He will also engage in coping processes which can divert resources
away from job activities and consequently reduce job performance.
The key element of the concept of emotion episodes is that these episodes
represent the ebb and flow of emotional experience over time. While each of
the events during the episode can be described in discrete terms, the episode
itself has a coherence and a set of features that suggest it should be treated
as a unit of analysis.
AFFECTIVE CYCLES:
EXOGENOUS AND ENDOGENOUS CAUSES
Earlier, we stated than any analysis of affect as either a dependent or
independent variable would require a full consideration of the fluctuation of
affect levels over time. We are now prepared to offer a more detailed, albeit
speculative, statement of our position.
We begin by suggesting that affect states can be described at two levels. First.
one can describe an immediate affective state at a general or "primary" level.
This description would refer to the state as being either positive or negative
and of being at a certain level of intensity. In some cases, affective states can
only be described at this primary level. A person's current mood, divorced as
it is from specific references, is the notable example. In other cases, when
specific emotions are involved, the affective state can be described at two levels,
a general description involving valence and intensity (corresponding to the
primary appraisal of emotion generation) and a secondary level which
references the specific emotion(s) involved. Even where specific emotions are
experienced, description and analysis at the primary level can still be useful. i
Primary affective states can and do fluctuate over time. Particular trends ]
in affect levels are the result of both endogenous and exogenous factors. 1
Endogenous factors which influence affect trends include affective dispositions, 1
established cycles of mood levels and chronic, situational circumstances of
Affccf ivc Ez~cnts Theoy 43
affective significance. Exogenous factors are affective events which serve as

"shocks" that disrupt the regularity of underlying affect patterns. They can be
either work related or nonwork related.
In most cases, these shocks will constitute events which generate an
emotional reaction consisting of both primary and secondary appraisal.
Consequently, the affective state can be described both in terms of the general
evaluative dimension and in terms of the specific emotion experienced.
However, in some cases events of "mild hedonic relevance" may elicit a change
in mood state without accompanying secondary appraisal and therefore
without the experience of a specific emotion.
Finally, affective "shocks" produce "after shocks." Here we are referring to
the emotion episodes we described earlier. To remind the reader, an emotion
episode refers to a series of emotional experiences precipitated by a single
emotional event. During the emotion episode the person is in a state of
continuous affective or emotional engagement and the series of after shocks
should continue to alter the normal affective pattern.
In Figures 8 and 9 we have tried to visually depict what we have been
suggesting. Figure 8 shows a hypothetical picture of a pattern of affect for a
single individual. Figure 9 shows ones subject's data from a study by Weiss,
Nicholas, and Daus (1993). Weiss et al. asked managers to provide self-reports
of affect levels 4 times a day over a 3-week period. Figure 9 shows the results
for one manager over 4 days, 16 total observations, presenting only the results
for self-reports of being in a pleasant or unpleasant mood state.
It can readily be seen that the collection of affect data over any period of
time allows for analysis of a number of parameters in affect trends, including
means and variances of the aggregate set as well as the frequency and severity
of "affect spikes." These parameters can serve as both dependent and
independent variables in organizational analysis. For example, one might
examine the relationship between dispositional variables and the mean or
variance in affect states. Spikes serve as evidence for the occurrence of affective
events which disrupt more typical affect trends and, therefore, can be examined
in relation to these events. All of these parameters can be examined with regard
to their association with traditional organizational criteria.
However, focusing on parameters which summarize these data hides the
richness of the information inherent in data examined over time, its rhythm
and its deviations from an underlying rhythm. Oscillating patterns or cycles
of mood have been examined in nonwork settings by Larsen and Kasimatis
(1990) and McFarlane, Martin, and Williams (1988) among others. Such cycles
can be examined in terms of the amplitude of response as well as the frequency
or period of the cycle. Larsen and Kasimatis (1990), reminiscent of Hersey
(193% showed that the extent to which a 7-day cycle described mood
fluctuations was predicted by individual personality factors, with introverts
showing a more cyclical pattern than extroverts. Studies of mood cycles within
P-A
E-A
z-A
C-A
P-41
c-41
z - 41
-
1 41
P-M
E-M
Z-M
C-M
P-1
E-1
z-1
C-1
days, requiring multiple measurements throughout the day, are less frequent
and apart from Hersey's speculations, formal analyses of mood cycles with
data collected in work settings do not exist.
As an example of the possible advantage of looking at patterns of affect,
we offer the study of the relationship between affect and performance. While
reported associations of job satisfaction with performance are basically
negligible, such analyses involve correlating aggregrate performance with
"aggregate" satisfaction. In our opinion, examining the correspondence of
affect patterns over time and performance patterns over time would be a more
productive approach.
Cycles represent the endogenous components of the mood trends while
exogenous components are represented by deviations from the cycles. One can
focus on the stable patterns and treat the exogenous events as error, as do
researchers searching for mood cycles, or one can detrend the data to examine
effects of exogneous shocks. Each approach has a role to play in organizational
analyses of affect.
While we think that the examination of the trends in affect is of real
significance we do not want to misrepresent our position. To us, the affect
trends are simply the history of affective experience over a period of time. As
a documentation of that history, it may have predictive significance. However,
the real action lies with the affective state at any particular time.
AFFECT AND SATISFACTION JUDGMENTS

A basic premise of this paper is that affective experiences at work influence
overall judgments about satisfaction with one's job and do so independent of
the influences of feature assessments. This position is consistent with more basic
research on the affective and "cognitive" influences on general attitudes. In two
separate studies, Abelson, Kinder, Peters, and Fiske (1982) asked survey
participants to describe Presidential candidates in affective and trait terms.
More specifically, participants were asked whether a candidate "has ever made
you feel (angry, hopeful, etc.)" and also to evaluate that candidate in terms :
of specific traits (honest, smart, etc.). Regression analyses were conducted to 1
i
examine the independent influence of both affective and trait ratings on overall 1
evaluations of the candidates. In both studies the affect ratings consistently I
predicted the overall evaluations independent of the trait ratings. The trait
ratings in fact were less useful. I
More recently, Breckler and Wiggens (1989) reported two studies attempting
to validate the distinction between affective and what they referred to as
"evaluative" or belief influences on overall attitudes. In the first study students
were given a list of objects or activities (e.g., computers, blood donations).
Using scales developed for the study they were asked to indicate their beliefs
Affective Ezjents Theory 47
about the objectslactivities and how each makes them feel. These were then
correlated with global ratings of each object/ activity. Partial correlational
analyses indicated that the affective and belief components had independent
influences on the overall attitudes. The second study replicated the first and
also showed that the affective component better predicted behavior then did
the belief component.
Taking the distinction between affective and belief components a step
farther, Edwards (1990) showed that attitudes can develop out of both affective
experiences and information about the attitudinal object. Moreover, the
effectiveness of persuasion attempts depends on whether the nature of the
attempt is consistent with the original mode of attitude formation. In addition,
Miller and Tesser (1986) argued that instrumental behaviors are driven by the
belief component and consummatory behaviors by the affective component.
They demonstrated that by getting people to focus on one of the components
they could influence subsequent behaviors involving the attitudinal object.
All four papers proceeded from the basic premise that attitudes are general
evaluations which are influenced by both affective experiences with the object
and sets of more abstract beliefs about the object. While these are often referred
to as the affective and cognitive components of attitudes, Breckler and Wiggins
(1989) correctly point out that because cognitive processes are an important
part of affective reactions it is better to make a distinction between those
influences which represent emotional experiences with the object and those
influences which relate to sets of beliefs about characteristics of the object.
Taken together, these studies demonstrate the usefulness of making that
distinction.
The relevance of this work to the nature of job satisfaction is easy to see.
The affective component of job satisfaction reflects the recall of affective
episodes on the job and the belief component represents evaluations of the
job in terms of its features, as operationalized in the various versions of
cognitive/judgment theories.
With regard to job satisfaction, the importance of the belief component has
been frequently demonstrated. The contribution of the affective component
has been less well established. T o help rectify this situation, Weiss, Nicholas,
and Daus (1993) had a sample of managerial/ professional employees complete
diaries in which they reported their mood levels four times daily over a 3-week
period. At the end of the three weeks and independent of the diary procedure,
participants completed a questionnaire which included a 5-item overall job
satisfaction scale as well as a VIE based belief assessment containing questions
about the instrumentality of the job for receiving each of 10 outcomes and
the valence of those outcomes.
While the intrusiveness of the data collection procedure limited participation
to only 24 subjects, the results clearly support the relevance of affect as a
predictor of satisfaction and the relative independence of affective and belief
based influences. Based on considerations described earlier, four different

operationalizations of the mood self-reports were used: pleasantness,
activation, positive affect, and negative affect. Average mood levels were
computed (4 times daily over 16 days) and correlated with overall satisfaction.
Three of the four correlations were statistically significant ( r = .66 for
pleasantness, r = .47 for positive affect, r = -.41 for negative affect) while
the fourth ( r = .36 for activation) approached significance.
Perhaps more importantly, regression analyses in which overall satisfaction
was regressed on mood and VIE beliefs together demonstrated that both
components had significant, independent influences on job satisfaction.
Without getting into the question of which component is more important, the
answer to which is probably "it depends," it is clear that with regard to
judgments of job satisfaction both affective experiences and more abstract
belief systems play a role.
The distinction between mood and emotion further complicates the analysis
of affective influences on job satisfaction. We have argued, based upon Morris
(1989), that moods are general affective states disassociated from particular
events. They can arise as residual states after events or in response to the recall
of previous emotional events. They can also arise from factors like the weather,
having nothing to do with particular events. While, as in Weiss, Nicholas, and
Daus (1993), one can chart the mood levels of individuals over the course of
a day, it is more likely that specific events and not general mood levels are
stored and recalled. Consequently, it may be that average mood level over a
period of time predicts satisfaction (as in Weiss et al., 1993) because it is an
indicant, however imprecise, of the frequency and/or intensity of affective
events. It may also be the case that the average mood level predicts satisfaction
because both share affective dispositional variance.
These positions are not incompatible and both suggest that part of a
relationship between average mood level and satisfaction is spurious. If so, a
better analysis of affective events and subsequent job satisfaction would focus
on the occurrence of the events themselves and the affective reactions they
generate or, at the very least, focus on significant changes in affect levels as
indicants of affective events.
Not all of the relationship is likely to be spurious. To begin with, mood may
color the interpretation of events. As such, being in a negative mood may result
in a neutral event being interpreted as negative. More directly, a person's mood
may trigger an actual event of affective relevance (e.g., an argument, a
reprimand).
As certain as the dual influences of affect and belief on attitudes appear to
be, the evidence for these influences does not speak to the basic question of
whether attitudes are formed stored, revised, and recalled or whether they are
constructed on demand. With regard to job satisfaction, is the evaluative
judgment continuously revised by experience and therefore in place when
AffectiveEvents Theory 49
required by the satisfaction questionnaire or does the questionnaire instigate

an attitude construction process and a new judgment?
Hastie and Parke (1986) have made the distinction between "on-line" and
memory based judgments in person perception and this distinction has
relevance here. With "on-line" judgments, evaluations are made at the time
of exposure to the events or information relevant to the judgment. Later, a
stored judgment is retrieved. With memory based judgments, the judgment is
made later and based upon retrieval of representations of the initial events or
information.
Wilson and Hodges (1992) have provided an important discussion of this
topic with regard to attitudes. They begin their discussion with what they refer
to as the traditional viewpoint, a viewpoint which states that attitudes are stable
and persistent evaluations. According to the traditional view, when people are
asked to report their attitudes toward some object, their job, their supervisor,
and so forth, they open a "mental file containing their evaluation" and report
it. Presumably, this evaluation is the result of on-line judgment processes.
The alternative position, which they advocate, states that people often
construct their attitudes at the time the attitude is called for. In constructing
these attitudes they consult a large "database" of relevant information
including, presumably, affective experiences with the attitudinal object and
belief based information. When constructing these attitudes the full database
is never tapped. Rather, at any given time, people generate their attitudes from
a subset of the "database" and the subset they use can itself be influenced by
current contextual factors. Consequently, attitudes can be quite unstable
because they are influenced by situational factors which'determine which
information is utilized.
Wilson and Hodges (1992) offer a good amount of evidence to support their
alternative position, focusing on the extensive evidence on contextual effects
on attitudes. Of particular importance is research showing that current mood
state can influence attitudinal reports. From this literature it seems quite clear
that the mood of a person at the time satisfaction is being assessed can also
influence responses to satisfaction scales, independent of the history of affect
or the nature of belief systems. Recently, Brief, Butcher, and Roberson (1995)
rafher dramatically demonstrated this phenomenon. Employees filled out an
attitude survey in small groups. In some groups, prior to completing the survey,
employees were given cookies, soft drinks and an inexpensive gift, a
manipulation designed to enhance positive mood. Employees in other groups
were simply asked to complete the survey. Reported job satisfaction was higher
in the positive mood groups than in control groups.
Wilson and Hodges (1992) go on to say that not all attitudes are constructed
at the time of attitude evaluation. Some are, in fact, stored and recalled. Wilson
and Hodges offer a number of suggestions about which attitudes are likely
to be constructed and which are likely to be stored but perhaps the most well
worked explanation is offered by Fazio (1986). Fazio describes attitudes as

differing in "accessibility" which is simply the strength of the association
between an object and an evaluative response. Attitudes gain in accessibility
when then the association is strengthened by repeated calls for the evaluation
or when the attitude has been formed by direct experience with the object.
Given the validity of Fazio's position, one could argue that attitudes formed
on the basis of personal experience would be less influenced by contextual factors
such as current mood. One could further argue that attitudes toward certain
job facets, facets that have to do with "concrete" or tangible aspects of one's
job, like coworkers or supervisors, are more likely to be formed by personal
experience and are therefore less likely to be affected by context. More abstract
features, security, trust, career development, as examples, are more likely to
result from indirect informational influences and would be more affected by
context. The study by Brief, Butcher, and Roberson (1995) examined only
overall satisfaction and therefore does not allow for a test of this proposition.
Finally, if job satisfaction is at least partially constructed on demand one
could reasonably ask about the way affective experiences are used in the
judgment process. The nature of the storage and recall process is unclear. When
attitude judgments are made, are affectively meaningful events somehow
counted? Are they averaged over events with different intensities? If averaged,
are they weighted by recency?
Very little research has been done on these questions. Diener and his
colleagues (Diener, Colvin, Pavot, & Allman, 1991; Diener, Sandvick, & Pavot,
1990) have shown that the Subjective Well Being, a general judgment about
satisfaction with one's life, depends more on the frequency of positive events
than on their intensity. Diener speculates about a number of processes which
lead to "psychic costs" following the experience of intense positive affect,
thereby dampening their influence as well as the influence of subsequent
positive experiences. Regardless of the processes involved, the suggestion that
the frequency rather than the intensity of positive experiences has a more
pronounced influence on satisfaction has implications for organizational
practice. It suggests that major but infrequent affectively meaningful events
such as recognition ceremonies or bonuses will be less important determinants
of overall satisfaction than will working in an environment which provides
daily, if only minor, positive experiences.
Affective experiences and object relevant beliefs together influence evaluative
judgments. This much we know. Basic information about the cognitive
processes involved remains speculative.
AFFECT AND PERFORMANCE

Perhaps the most disappointing aspect of decades of study on work attitudes
is the failure to locate the holy grail of satisfaction research, an effect of
satisfaction on performance. The typical study examining satisfaction-

performance relationships has taken the following form: satisfaction, assessed
as a bipolar construct with positive (satisfaction) and negative (dissatisfaction)
poles is measured at some arbitrary point in time and correlated with some
aggregate measure of performance. It is generally known that such studies have
produced basically nothing. Satisfaction shows negligible correlations with
performance (Iaffaldano & Muchinsky, 1985; Podsakoff & Williams, 1986) and
when correlations are found causal direction is ambiguous. Correlations with
indices of withdrawal (turnover, absenteeism, lateness) are perhaps more
consistent but still quite small (Hulin, 1991).
Over the years numerous attempts have been made to explain the
discrepancy between the intuitive appeal of the proposition and the absence
of supportive results. So, for example, Fisher (1980) and Fisher and Locke
(1992) have suggested that the lack of a relationship between satisfaction and
performance was due to the discrepancy between the "generality" of attitudes
and the "specificity" of most criterion indices. Solutions to this problem of
inconsistent generality involve either making the attitude more specific,
assessing "attitudes toward the act," or making the behavior more general,
assessing aggregated patterns of multiple behaviors.
Neither solution is particularly satisfying. The former solution essentially
changes the concept of job satisfaction in the service of theoretically
uninteresting predictability. The second is perhaps more interesting but runs
the risk of creating aggregations that have little practical meaning to
organizations, also in the service of predictability.
Interestingly, rather than trying to explain the absence of findings one might
ask why any relationship would be expected. What is not often said is that
in the form in which the key constructs are conceptualized and measured there
is very little reason to expect any relationship to begin with. With regard to
productivity, one is hard pressed to develop a rationale for why a worker's
overall evaluation of his or her job should in any way influence how hard he
or she works on the job, what strategies he or she employs when doing the
job or any of the other factors which affect task performance. With regard
to withdrawal, the behavioral processes are perhaps easier to accept, generally
Involving decisions in which job satisfaction is one factor entering into a
withdrawal decision. Even here, however, the decision process seems better
suited to turnover than absenteeism or lateness which appear to be more
spontaneous and less "thoughtful."
Our interest is, of course, affect and our objective in this last section is to
examine the conceptual relationships between affective states and job
performance, broadly defined. Generally, we will keep our reviews of the
literature brief, as good reviews exist elsewhere (see Morris, 1989, for the basic
literature on the effects of moods and Isen & Baron, 1991, for a more
organizationally focused review of the effects of positive affect). Our focus will
be more on the conceptual structure of these linkages, although describing these

linkages will, of necessity, involve references to specific studies.
Affect versus Attitude Driven Behaviors
To begin with, any understanding of the relationship between affect levels

and work behaviors must begin by drawing the distinction between affect
driven behaviors and attitudinally driven behaviors. Certain work behaviors
are direct responses to affective experiences. So, for example, mood influences
helping behaviors, information processing strategies and probability judgments
(Morris, 1989). Similarly, more overt negative emotional experiences are likely
to lead to specific coping responses (Lazarus, 1991a, 1991 b). The particular
pattern of responses can have important effects on work performance and these
effects are not mediated by any relationship between affective experiences and
satisfaction judgments. Thus, for affect driven behaviors, direct affect-
performance relationships without nonspurious satisfaction-performance
relationships would be expected. Further, because affect levels can fluctuate
we would expect these affect driven behaviors to be of a relatively short
duration and be high variability. This would suggest that any relationship
would be best captured by analyses which assess to congruence of patterns
of affect and performance.
Other work relevant behaviors are attitudinally driven. Attitude driven
behaviors are directly influenced by overall evaluations of one's job and
consequently any relationship between affect levels and these behaviors will
be mediated by job satisfaction. These are likely to be behaviors which result
from well considered decisions and specifically, those behaviors where the
overall evaluation of the job enters into that decision.
Hulin (1991) has suggested that discrete withdrawal behaviors like lateness,
absenteeism, turnover, and retirement are in fact all manifestations of an
underlying latent withdrawal propensity. In his view, the common finding of
low correlations among these behaviors is primarily due to attenuation
resulting from their low base rates. Working from this position, Hanisch and
Hulin (1990) attempted to show that retirement is one manifestation of this
underlying withdrawal construct. In fact, what they showed was that the four
withdrawal behaviors sorted into two factors, "work withdrawal" which
included indices of unfavorable work behaviors, lateness and absenteeism, and
"job withdrawal" which included turnover intentions and desire to retire. In
addition, regression analyses suggested that JDI factors like pay satisfaction
or work satisfaction were better predictors of job withdrawal then work
withdrawal. This last finding is consistent with a generally better association
between satisfaction and turnover than satisfaction and absenteeism (Hackett
& Guion, 1985). (Lateness has been too infrequently studied to draw any
conclusions.)
While this two factor solution and the JD1 correlations require replication,
the findings are clearly consistent with our distinction between affect driven
and attitude driven behaviors. Turnover and retirement are behaviors that
require a thoughtful decision and satisfaction, the overall evaluation of one's
current job situation, is likely to be one factor entering into that decision.
Lateness and turnover are less "considered," tend to be more spontaneous, and
are likely to be more a function of immediate affect levels.
The distinction between affect driven and attitude driven behaviors rests on
the assumption that affect levels have direct behavioral consequences. In the
next sections we point to the evidence with regard to the effects of emotions
and moods.
Affect Driven Behaviors: Emotions
Throughout this paper we have distinguished between affective states which

are unfocused, relatively mild and exist as background to our daily experiences
(mood) and more overt emotional reactions of greater intensity of which people
are aware (emotions). Discussion of the nature of affect driven behaviors will
be facilitated by maintaining this distinction. Partly this is due to the fact that
the emotion and mood literatures tend to be distinct and partly this is due
to the fact that the effects of affective states, particularly negative ones, may
depend on the degree of awareness of the state. Nevertheless, it is important
to note that this distinction can be pushed too far. Certainly, most researchers
from both traditions rarely worry about a difference between mo,od and
emotion and therefore manipulations can overlap. Further, as mentioned
earlier, overt emotional events may be one cause of longer lasting mood states.
Emotional reactions are often characterized as dysfunctional, disorganizing
states. We believe, following Lazarus (1991a), that this is a misconception.
Emotional responses are generally quite functional, but they are functional qua
the emotion. They d o not disorganize behavior as much as they reorganize
or redirect it. Emotions d o interupt ongoing behavior (Lazarus, 1991a;
Mandler, 1984). They have what Frijda (1993) refers to as "control precedence."
People in an emotional state tend to be controlled by that state, they tend to
be preoccuppied by the emotion and there is a persistance to behaviors designed
to deal with the emotion. This control precedence is particulary dramatic in
the case of the emotion episodes which we described earlier. From the
perspective of the previous behavior, this may seem disorganized and
disruptive. However, from the point of view of the emotional.problem it seems
less so. Emotions "organize" behavior around the demands of the preciptiating
situation and "disorganize" the activities that were disrupted.
For our purposes, we need to examine the effects of this "reorganization"
response on job performance. In our opinion, the analysis of the performance
implications of emotional states reduces to a distinction between two domains
j Affective Events Theory 55
of behavior. Behaviors in the job domain are those behaviors required to do

one's job. Behaviors in the emotion domain are those behaviors driven by the
emotional state. The only question of relevance to job performance is this:
"How are the behaviors in the emotion domain related to the behaviors in the
job domain?"
Three answers are apparent. First, behaviors in the emotion domain can
interfere with behaviors in the job domain. Emotional reactions can produce
responses incompatable with job demands or can use up cognitive resources
needed for job performance. Second, behaviors in the emotion domain can
facilitate job performance, perhaps by increasing arousal levels or by instigating
performance compatable behaviors (e.g., increased social behavior for a
salesperson). Third, the behaviors in the emotion domain may be unrelated
to those in the job domain. The key point, of course, is that behavior is not
performance. The performance implications of emotions will depend on the
match or mismatch between behaviors generated by the emotion or the
resources used by the emotion response process and the behaviors and
resources required by the task.
This being said, it is our contention that in most cases the emotion responses
will tend to be incompatable with behavior in the job domain, producing
performance decrements. Interestingly, these decrements are likely to occur as
a consequence of both positive and negative emotions, because the
management of both types of emotions will require resources which could be
used for task performance and both types of emotions are more likely to
produce job incompatable rather than job compatable behaviors. This
prediction is in direct contradiction to the inuitive prediction of a positive
performance effect for positive emotions and a negative effect for negative
emotions.
In addition, we suggest that the performance implications of negative
emotions will be more pronounced than those of positive emotions. Negative
emotions serve as signals that some state of affairs is problematical. The
sequence of activity involves primary appraisal, secondary appraisal and the
development of a coping strategy designed to correct the situation. These
activities are likely to be more extensively and continuously disruptive than
are the activities resulting from the appraisal of a positive state.
We would therefore predict greater decrements in performance for negative
emotional states than for positive states and certainly would be hard pressed
to predict increments in performance for either state except under those rare
circumstances where the behaviors in the emotion domain and job domain
are compatable. Certainly, predicting simple linear associations between
affective states (positive to negative) and performance (positive to negative)
seems overly simple. The asymmetry for the performance implications of
positive and negative emotional states is consistent with the research by Hersey
(1932) which we cited earlier.
The behavioral implications of negative emotional states are generally

discussed under the topic heading of coping strategies. It is not our desire to
review this extensive literature. We do, however, want to comment on the
important position offered by Lazarus and Folkman (Folkman & Lazarus,
1990; Lazarus, 1991a). Their theory identifies eight different coping strategies
in response to a negative emotional event. These eight can be further reduced
to two categories: problem focused coping and emotion focused coping.
Problem focused coping deals directly with the emotion eliciting situation,
through problem solving or confrontation. Emotion focused coping tries to
deal with the affective state rather than its cause by denial, personal control
of affect, social support, and so forth.
The choice of strategy has obvious performance implications but one should
not be misled into thinking that problem focused coping is more performance
efficient than is affect focused coping. Problem focused coping involves
planning to deal with the emotional situation. This can be consuming of time
and resources and can have negative effects on performance. On the other hand,
affect strategies which involve denial might be more efficient in the short run
because they allow the person to focus on the job.
However, we also need to make the distinction between short-term and long-
term solutions. If we assume that elimination of the problem is the most
performance effective strategy in the long run then strategies which avoid
problem solving are not long-term efficient. In any case, research on the
relationships between typical coping strategies and performance on jobs with
different behavioral demands would be a useful activity.
Affect Driven Behaviors: Moods
The effects of mood on memory, judgments and behaviors have been studied
quite extensively in the last decade. Research by Alice Isen has stimulated much
of this work and extensive reviews of the literature can be found in Isen and
Baron (1991) and Morris (1989). While Morris' book primarily reviews the
basic research on the topic, Isen and Baron present reviews of both the basic
and organizational literatures.
Before proceeding, it will be useful to again remember the distinction
between behavior and performance. Mood has well documented effects on such
things as memory, judgment, and social behaviors but the performance
implications of these effects depend upon task demands. So, for example, mood
appears to influence the depth of processing on cognitive tasks (Sinclair &
Mark, 1992). The effect on performance will depend on the processing
requirements of the job.
Three additional factors complicate efforts to draws definitive implications
for organizational behavior from the basic literature on moods. First, mood
effects themselves are complex. Effects are often inconsistent and researchers
have most recently reduced efforts t o document main effects and instead have
searched for moderators and theoretical explanations that provide coherent
organization to apparent inconsistences. Consequently, we will provide an
overview of both basic findings and theoretical perspectives.
Second, the effects of mood are not always symmetric (Morris, 1989; Taylor,
1991). In the basic literature, much of the research has independently
manipulated positive mood or negative mood and then compared the mood
induced group t o a control group not experiencing the mood manipulation.
Even where both positive and negative moods are manipulated in the same
study, results have been generally interpreted independently for the effects of
positive mood, compared t o a neutral mood, o r negative mood, compared to
a neutral mood. These paradigms can hide the fact that when being in a positive
mood has effect A, being in a negative mood does not necessarily have an effect
opposite t o A. For example, being in a positive mood often (but not always)
increases helping behavior when compared to a control condition of a neutral
mood (Isen & Baron, 1991). However, being in a negative mood can also
increase helping behavior, again when compared to a neutral condition
(Morris, 1989).
Third, positive mood effects are almost entirely demonstrated by way of
experimental studies where mood is manipulated with events of minor
positive hedonic relevance. Isen's work (Isen, 1984; Isen & Baron, 1991) best
illustrates this approach. People find money in telephone booths, they are
given small prizes, they are shown funny movies o r listen t o pleasant music.
Negative mood effects, on the other hand, are demonstrated in two distinct
ways. First, events of minor negative hedonic relevance are manipulated.
People are told they have failed some task o r see a depressing movie o r are
asked t o read sad stories. These kinds of studies parallel the positive affect
studies and their results can be more o r less compared t o examine the
differential effects of positive and negative moods. Negative mood effects
are also demonstrated by comparing chronically, clinically depressed people
with normals o r sometimes by comparing less severe but still chronically
d ~ s p h o r i cpeople with normals (Conway & Giannopoulos, 1993). These
studies raise all of the traditional concerns of causal direction and
s~uriousnessas well as new issues of the effects of mood intensity. We believe
that the results of studies where negative mood is manipulated are of greater
relevance t o work issues than are results drawn from research with the
chronically depressed.
The effects of mood have generally been organized into four categories:
mood effects on memory, mood effects on evaluative judgements, mood effects
On Processing strategies, and mood effects on social behaviors. All have
relevance to job performance and we will discuss each in turn.
Mood and Memory
One of the most extensively researched topics is the effect of mood on

memory. Most of the research has been stimulated by Bower's (1981)
integration of mood and emotion with an associative network model of
memory. In this model emotions, like ideas and events, are represented as nodes
in a network of associations or linkages. Whenever a node is stimulated, for
whatever reason, activation spreads to connected nodes. If the activation level
crosses some threshold the idea represented by the connected node is brought
into consciousness (remembered). A number of predictions for the effects of
mood on memory can be generated by this.mode1but the two that have received
the most attention have been the predictions of "state dependent memory" and
"mood congruent memory."
State dependent memory refers to a facilitation of memory when the
psychological or physiological state at the time of learning matches the state
at the time of recall. So, for example, if you are in a negative mood when
you learned some material your later recall of that material should be facilitated
when you are again in a negative mood. This prediction derives from the idea
that the material and the mood are associated at the time of learning. Activation
of the mood at time of recall spreads to the learned material increasing the
likelihood of the material reaching the activation threshold and thereby being
brought into awareness. The hedonic tone of the learned material (positive,
neutral, or negative) is of no relevance to the facilitation of recall.
Initial research by Bower and his associates seemed to support this position
( ~ o w e r ,1981). Subsequent research has been neither strong nor consistent
(Clore, Schwarz, & Conway, 1994; Morris, 1989) leading even Bower (Bower
& Mayer, 1985) to question its robustness. (But, see Eich, 1995, for a discussion
of some conditions which may facilitate the state dependent memory effect.)
Mood congruence effects are demonstrated when memory is facilitated by
a match between the hedonic tone of the material being recalled and the mood
at the time of recall. So, for example, one might ask whether being in a good
mood facilitates the recall of positive as opposed to negative material. This
issue of recall congruence has been considered important to mood researchers
because the phenomenon has implications for the effects of mood on judgments
and behaviors that rely on the recall of information. Consequently, a good
amount of research has accumulated on the topic.
Overall, mood congruence finds more support than does state dependent
memory, but the results are still not entirely straightforward. Clore, Schwarz,
and Conway (1994) have most recently reviewed the literature on mood
congruent recall and conclude that results support a congruence position, but
not as would be predicted from an associative network model. More
specifically, they argue that while positive moods facilitate the recall of positive
materials and inhibit the recall of negative materials negative moods inhibit
the recall of positive material but do not facilitate the recall of negative material.
Clore et al. suggest that this asymmetry may be due to a mood management
or mood repair process, whereby people in a negative mood deliberately avoid
the recall of negative material in order to reduce their negative mood. However,
as Clore et al. (1994), Blaney (1986), and Morris (1989) all point out, this mood
repair process also suggests that people in a negative mood will deliberately
recall positive material as well as repress negative material and this is not
generally found. In any case, this issue of incongruence in the way positive
and negative moods affect the recall of positive and negative material does
not invalidate the basic finding; people in a positive mood recall more positive
items from memory than negative items (recall of positive items is facilitated
and negative items inhibited) and people in a negative mood recall more
negative items than positive items (recall of positive items is inhibited).
Mood and Evaluative Judgments
Given the effects of mood on memory, one would naturally expect to find
a biasing effect of mood on memory based judgments. In fact, this bias has
been repeatedly demonstrated. So, for example, mood influences the
evaluation of the positiveness or negativeness of events, the evaluation of
neutral objects, the evaluation of other people and the perceived likelihood
of positive or negative events occurring (Morris, 1989). All of these influences
are in a mood congruent direction.
A good organizational example of this biasing effect is provided by the Brief,
Butcher, and Roberson (1995) paper which we discussed earlier. They had two
groups of employees complete a traditional job satisfaction questionnaire.
Recall that prior to filling out the survey, some employees were placed in a
positive mood while the remainder of the employees were not given a positive
mood induction and served as the control group. As expected, employees in
the positive mood reported higher job satisfaction than did employees in the
control group.
The biasing effect of mood on evaluative judgments would appear to be a
direct outcome of mood congruent memory effects. That is, when people are
called upon to make judgments of evaluation, event probabilities, and so forth,
they search their memories for information relevant to the judgment. If the
mood they are in influences what they recall, as it does, then it should also
influence the judgments that are based on these recollections.
AS compelling as the memory based explanation is for these effects of mood
on judgment, it does not appear to account for the phenomenon. Two findings,
In particular, cast doubt on the memory based explanation. Johnson and
Tversky (1983) had subjects read newspaper stories about deaths from different
causes (illness, murder, fire) and then complete a questionnaire asking about
the risks of dying by various causes. As intended, the newspaper articles
produced a negative affective state in the subjects and subjects in these negative
states reported greater concerns about dying from various causes than did
controls. Of particular interest is the finding that the concerns were unrelated
to the type of death that produced the original negative mood state. That is,
people whose negative mood was caused by reading about someone dying from
leukemia showed no greater concern about dying from illness than about dying
from an accident. This finding, as Johnson and Tversky pointed out, is
inconsistent with a memory based process because such a process would predict
that the mood induction information would be most likely to cue memories
related to that event.
Schwarz and Clore (1983) asked people in positive and negative mood
conditions to respond to a life satisfaction questionnaire. Ordinarily, as
predicted from any memory based explanation, one would expect that current
mood would influence reports of life satisfaction. However, Schwarz and Clore
also provided some subjects with an explanation for why they were in the mood
they were in. Their findings clearly showed that current mood only influenced
life satisfaction evaluations when subjects were not given a reason for their
mood state. This finding, replicated on numerous occasions (Clore, Schwarz,
& Conway, 1994), is difficult for any memory based explanation since these
explanations posit mood as a memory cue regardless of any attributional
information which might be available.
How is the mood effect on evaluative judgments to be explained if not by
biased recall from memory? An alternative position, gaining popularity, is the
"mood as information" position advocated by Schwarz and Clore (1988; Clore,
1992). The basic idea of this position is that the mood you are in is a piece
of information you use to tell you how you feel about the object being judged.
Essentially, you misattribute your affective state to the object or issue being
evaluated and your judgment is biased accordingly. Let us illustrate this
position by referring back to the paper by Brief, Butcher, and Roberson (1995).
The reader will recall that employees placed in a positive mood at the time
of completing a job satisfaction questionnaire reported higher job satisfaction
than did control subjects. A memory based explanation for these results would
suggest that the positive affect facilitated the recall of positive information used
in making the satisfaction judgment. The mood as information position would
suggest that the employees misattributed their current mood to the topic of
the questionnaire, thereby assuming they felt better about their jobs than they
would have in a neutral mood state. The mood as information position would
also predict that the mood biasing effects on satisfaction can be eliminated
by providing attributional information when the survey is administered.
The mood as information position suggests conditions when the biasing of
effects of mood should be reduced. The existence of these moderating
conditions does not, however, reduce the importance of mood on
organizationally relevant evaluative judgments. We have previously defined
mood as affect disconnected from its causes. In the Schwarz and Clore
paradigm subjects are provided with clear, but false, explanations for their
moods. This does not imply that in natural contexts such clear attributional
factors will be readily available or that people in particular moods will be
actively searching for these attributions. Consequently, it seems likely that the
moderating condition of clear mood explanations is a low probability event
in work contexts and therefore main effects of moods on evaluative judgments
are more likely than not. Because so many organizational behaviors involve
judgments and decisions based on evaluations of the likelihood of positive and
negative outcomes the impact of mood states seems clear.
Mood and Processing Strategy
Sinclair and Mark (1992) have recently organized the diverse effects of mood
on judgment in terms of the effect of mood on processing strategy. They argue
that the findings consistently point to the conclusion that people in a positive
mood are more likely to engage in simplified, heuristic processing strategies
when making judgments and decisions while people in a negative mood are
more likely to engage in systematic processing. The evidence they marshal1 for
this general phenomenon is quite convincing . For example, in persuasion
studies people in a positive mood are more influenced by peripheral cues such
as source attractiveness while people in a negative mood are more influenced
by central cues like message quality (Worth & Mackie, 1987). Such a difference
in cue use is generally attributed to differences in depth of processing (Eagly
& Chaiken, 1993). In addition, when making judgments about others, people
in positive moods tend to rely more on stereotyping information while people
in negative moods rely more on individuating information (Bodenhausen,
1993). Positive mood appears to increase halo among judgments on different
dimensions when compared to neutral and negative moods (Sinclair, 1988).
Finally, when compared to controls, people in a positive mood use fewer
categories to group objects (Isen & Daubman, 1984).
In aggregate, these findings seem to converge on a general processing
difference between people in positive and negative moods. People in a positive
mood are less likely to engage in effortful, systematic processing than are people
in negative moods. This consistent use of heuristic processing can, of course,
be dysfunctional when tasks demand more effortful processing. For example,
Sinclair and Mark (1992) found that statistics students placed in a positive
mood were less accurate in their estimation of correlation coefficients from
scatterplots than were students placed in a negative mood.
Sinclair and Mark (1992) examine the potential explanations for this general
finding of processing differences and conclude that a combination of two seem
reasonable. The first explanation involves mood maintenance or mood repair
Processes. People in a positive mood may avoid systematic processing because
engaging in effortful cognitive processing may eliminate the positive mood.

People in a negative mood, on the other hand, may be more inclined to engage
in systematic processing as a distraction to help eliminate the negative mood.
The second and possibly coexisting explanation involves the mood as
information explanation described earlier. This explanation assumes that
current mood states are used as cues for judgments about the satisfactoriness
of situations. People in negative moods judge the situation as somehow
unsatisfactory, are motivated to rectify the situation and engage in the cognitive
activity necessary to accomplish that goal. People in positive moods judge the
situation as more satisfactory, less problematic, and therefore feel that
systematic processing is unnecessary.
We would be remiss if we did not comment on a line of research which
appears, on the surface, to be inconsistent with the conclusions of Sinclair and
Mark (1992). Ellis and Ashbrook (1988) have suggested that chronically
depressed individuals engage in more self-focused attention and negative
thinking than do normals. This, in turn, reduces the cognitive processes
available for problem solving producing a performance decrement on tasks
that require extensive cognitive processing.
Recently, Conway and Giannopoulos (1993) provided evidence to support
the Ellis and Ashbrook position. They provided depressed and non depressed
subjects with information about 70 different jobs along five dimensions and
asked the subjects to provide overall evaluations of the jobs. Using a regression
based policy capturing procedure they found that the depressed subjects used
fewer dimensions in generating their evaluations than did the controls.
Clearly, this data and the Ellis and Ashbrook position generally need to be
reconciled with the evidence provided by Sinclair and Mark. One obvious
difference is that Ellis and Ashbrook focus on the effects of chronic depression,
a negative affective state of longer duration and greater severity than the mood
manipulations used in studies cited by Sinclair and Mark. Whether this
accounts for differences in results remains to be determined. However, if the
effects of negative mood on processing strategy are curvilinear then questions
of appropriate levels for generalization to organizational contexts become
significant.
Recently, Staw and Barsade (1993) examined the effects of affective traits
on managerial processing strategies. First year MBA students participated in
an assessment center which included an in basket exercise measuring decision-
making strategies (use of information, accuracy of judgments, etc.). Individuals
assessed as having either high, medium, or low trait positive affect were
compared. Results indicated that compared to subjects with medium or high
trait positive affect, subjects with low trait positive affect used less data,
requested less information and were less accurate in their judgments. Staw and
Barsade's results are more consistent with the Ellis and Ashbrook position than
they are with the Sinclair and Mark position. However, it must be noted that
Staw and Barsade focused on trait affect levels rather than state affect levels.
It is worth investigating whether trait based affect differences evoke different
cognitive processes than state changes in normals. In any case, the Staw and
Barsade findings in addition to the apparent conflict between the Ellis and
Ashbrook and Sinclair and Mark positions provides further evidence of the
complexity of mood predictions.
Mood and Behavior
Helping behavior is the most frequently studied behavioral consequence of

mood, stemming directly from the work by Isen and her colleagues (Isen, 1984;
Isen & Baron, 1991). Given the potential relevance of helping to such important
organizational relevant variables as citizenship behavior and customer service,
it is not surprising that Isen's work has received a good amount of attention
in the organizational literature (see, for example, George, 1991; George & Brief,
1994). The very general finding of Isen's work on this topic is that positive
mood enhances helping behavior and cooperation while reducing
aggressiveness. Both Isen (1984) and Isen and Baron (199 1) review considerable
laboratory evidence indicating that being in a positive mood makes people
more generous, helpful, and cooperative, when compared to neutral mood
controls. While in a good mood people report greater liking for others and
evaluate people more favorably.
Moreover, it appears that it is easier to resolve conflicts among people in
a positive mood. For example, in the context of organizational negotiations,
people in a good mood are more likely to reach "win-win" integrative
bargaining solutions (Carnevale & Isen, 1986) and are more likely to make
cooperative concessions during negotiations (Baron, 1990).
The one caveat to the general finding of increased helping and cooperation
is that the effect is lost if people believe that helping may interfere with the
mood (Isen & Baron, 1991). Isen explains this exception to the general finding
by invoking two explanations. First, there is a basic tendency for positive
moods to increase helping. Second, there is a coexisting desire to maintain
a positive mood. When the helping interferes with the second desire it is
avoided.
As often turns out to be the case, the effects of negative mood are more
complicated. Because Isen's work compares positive mood subjects to controls,
the findings do not enlighten us as to the effects of negative mood. However,
one might expect that if being in a positive mood increased helping and
cooperation being in a negative mood would decrease it. This turns out not
to be the case. In fact there is extensive evidence to indicate that being in a
negative mood also can serve to increase helping behavior (Carlson & Miller,
1987). The most likely explanation for this effect involves self-regulatory
Processes wherein people in negative moods engage in helping behavior to
eliminate the negative mood (Morris, 1989). When people are led to believe
that their helping will not reduce their moods they do not help (Manucia,
Baumann, & Cialdini, 1984).
Although both Isen and Baron (1991) and George and Brief (1994) provide
thoughtful discussions of the relevance of the mood and helping literature to
organizational behavior not much empirical research has been conducted.
George (1991) found that a state measure of Positive Affect did correlate with
a measure of helping and also with a measure of customer service behavior.
No other quality empirical studies exist. However, given the complexity of the
basic findings, the moderators of the effect of positive mood on helping, and
the possibility that negative mood can also increase helping behaviors under
certain circumstances, predictions in organizational settings will not be
straightforward.
Given that being in a positive mood increases the estimated probabilities
of the occurrence of positive events and being in a negative mood increases
the estimated probabilities of the occurrence of negative events (Wright &
Bower, 1992) an effect of mood on behaviors based on probability estimation
would be expected. In fact, the relationship between mood and risk taking has i
been well researched and consistent effects are found, but, as before, these 1
effects are not what one might initially expect. In fact, it is with risk taking
that the push and pull of cognitive processing and mood maintenance seem
to show up most clearly. In the classic study on the topic, Isen and Patrick
(1983) showed that subjects given a hypothetical risk situation were willing to
risk more than controls, but when they were placed in a real situation where
their own resources were put at real risk they engaged is less risky behavior.
Later Isen, Nygren, and Ashby (1988) showed that people in a positive mood
expressed higher negative utilities for loss than did controls, apparently
overcoming any event probability differences when making decisions about
risky outcomes.
Throughout this section, we have emphasized the complex nature of mood
effects on judgments, processing strategies, and behaviors. Nothing in this area
appears to be straightforward. However, the complexity of findings should not
be confused with an absence of findings. Mood effects, while complex and
interactive, are also consistent and pronounced.
Overall, it seems that mood effects are a complex interplay of both cognitive
and motivational factors. On the cognitive side, mood effects on memory,
attributions and processing strategies play a role. On the motivational side,
mood repair (for negative mood states) and mood maintenance (for positive
mood states) also influence responses. A particular prediction in any situation
depends on the mood state, the task, and the aforementioned cognitive and
motivational processes.
Affective Events Theory
FINAL COMMENTS
What we have attempted to do in this paper is to flesh out a structure for an
event based approach to affective experiences at work and to fill in some of
the details by referring to the basic literature on emotions and moods. Along
the way, we have contrasted our position with some more traditional positions
on similar topics. In some cases we offer our position as a complement to those
traditional positions, in other cases as an alternative. In all cases we hope to
encourage readers to rethink some issues they may have taken for granted.
Because we summarized our position early in the paper, we think it
unnecessary to do so again. Nevertheless, there are some key points and
implications which need to be reiterated.
To begin with, job satisfaction is not affect and it is time we stopped saying
it is. Over the years we have developed a construct called job satisfaction. This
construct, as defined by its most popular operations, is the overall evaluation
one has towards one's job. There may be differences of opinion about the
usefulness of the construct but the construct exists as what it is and it is not
affect.
Second, affective experiences in their traditional forms as emotions and
moods are potentially important aspects of work experience. These affective
states influence performance and job satisfaction but their performance
implications are, by and large, independent of their relationships with
satisfaction. It is the failure to realize this point, in conjunction with the
aforementioned confusion of affect and satisfaction, that has led to the dismal
record of satisfaction-performance research.
Third, affective states can and do fluctuate over time and the performance
implications of affect depend on affect states at particular times. Satisfaction
is also influenced by current affect levels, along with affect histories.
Consequently, time of assessment cannot be arbitrary and patterns of change
over time become primary dependent variables. When change becomes our
focus of attention, we have to modify the structure of our causal explanations
as a consequence. Explanations for change cannot be found in stable, steady
characteristics of people or situations. Instead they are found in discrete events.
Take the orbits of the planets as an example. The continued stable pattern
of planetary movement is "explained" in terms of the functional relationships
of the stable characteristics of the system (e.g., the mass of the planets and
the sun, etc.). How would we explain a planet pulling out of its orbit? Any
explanation would require the addition of a discrete event, like a meteor.
So too with our organizational variables. We must understand when we are
interested in stable patterns and search for functional relationships among state
variables, and when we are looking at changes and search for precipitating
events. In this paper, our focus has been on changes in affect levels, their causes
in events and their consequences in changes in performance. However, we have
also recognized that there are endogenous elements to the levels and patterns
of affective states which are most efficiently explained in terms of stable
attributes.
Finally, it is perhaps somewhat injudicious for us to admit that as we
developed our position our thoughts kept returning to the work of Frederick
Herzberg. We say injudicious because we are well aware of the way Herzberg's
theory is treated in current I / O Psychology. About the only thing positive
anyone has to say about his work is that it focused attention on intrinsic task
characteristics as influences on motivation and satisfaction. Beyond this, the
ideas remain only of "historical interest" (see, for example, Dipboye, Smith,
& Howell, 1994).
Certainly, we did not write this paper to defend Herzberg. We are well aware
of the methodological flaws in the research and conceptual problems with the
theory as a whole. Yet, in our opinion, Herzberg began with some fundamental
assumptions, the validity of which can be evaluated independent of the theory
in its entirety. People react to the events of their work lives. These events drive
their immediate affective states and these states can vary over time. Some events
are positive, some events are negative, some features are more likely to generate
positive events some features are more likely to generate negative events. These
were Herzberg's basic assumptions and these ideas inform our own position.
ACKNOWLEDGMENT
We wish to thank Seymour Adler, Arthur Brief, Rebecca Henry, and Carolyn
Jagacinski for their helpful comments on earlier versions of this manuscript.
REFERENCES
Abelson, R.P., Kinder, D.R., Peters, M.D., & Fiske, S.T. (1982). Affective and semantic
components in political person perception. Journal of Personality and Social Psychology,
42,6 19-630.
Adler, S., Skov, R.B., & Salvemini, N.J. (1985). Job characteristics and job satisfaction: When
cause becomes consequence. Organizational Behavior and Human Decision Processes, 35,
266-278.
Almagor, M., & Ben-Porath, Y.S. (1989). The two-factor model of self-reported mood: A cross-
cultural replication. Journal of Personality Assessment, 53, 10-21.
Arvey, R.D., Bouchard, T.J., Segal, N.L., & Abraham, L.M. (1989). Job satisfaction:
Environmental and genetic components. Journal of Applied Psychology, 74, 187-192.
Arvey, R.D., Carter, G.W., & Buerkley, D.K. (1991). Job satisfaction: Dispositional and
situational influences. In G.L. Cooper & I.T. Robertson (Eds.), International review of
industrial and organizationalpsychology (Vol. 6, pp. 359-383). Chichester, England: John
Wiley & Sons.
Arvey, R.D., McCall, B.P., Bouchard, T.J., & Taubman, P. (1994). Genetic influences on job
satisfaction and work values. Personality and Individual Differences, 17, 21-33.
Baron, R.A. (1976). The reduction of human aggression: A field study of the influence of
incompatible responses. Journal of Applied Social Psychology, 6, 260-274.
Baron, R.A. (1984). Reducing organizational conflict: An incompatible response approach.
Journal of Applied Psychology, 69, 272-279.
Baron, R.A. (1990). Environmentally-induced positive affect: It's impact on self-efficacy, task
performance, negotiation, and conflict. Journal of Applied Social Psychology, 20, 368-
384.
Bell, P.A., Fisher, J.D., & Loomis, R.J. (1978). Environmenfal psychology. Philadelphia, PA:
Saunders.
Bell, P.A, Garnand, D.B., & Heath, D. (1984). Effects of ambient temperature and seating
arrangement of personal and environmental evaluations. Journal of General Psychology,
110, 197-200.
Berkowitz, L. (1989). Frustration-aggression hypothesis: Examination and reformulation.
Psychological Bullefin, 106, 59-73.
Bittle, M.L., & Hausenstein, N.M.A. (1990, August). Deferminanfsofjobsafisfacfion:Socialcues,
ask sfrucfure.andpredisposi~ion.Paper presented at the annual meeting of the American
Psychological Association, New Orleans, LA.
Blaney, P.H. (1986). Affect and memory: A review. Psychological Bulletin, 99, 229-246.
Bodenhausen, G.V. (1993). Emotions, arousal, and stereotypic judgments: A heuristic model of
affect and stereotyping. In D.M. Mackie & D.L. Hamilton (Eds.), Aflecf, cognifion, and
s~ereotyping:In~eracriveprocesses in group perception ( pp. 13-37). San Diego: Academic
Press.
Bolger, N. (1990). Coping as a personality process. The Journal of Personality and Social
Psychology, 59, 525-537.
Bolger, N., & Schilling, E.A. (1991). Personality and the problems of everyday life: The role of
neuroticism in exposure and reactivity to daily stressors. Journal of Personality, 59, 355-
386.
Bouchard, T.J., Jr., Arvey, R.D., Keller, L.M., & Segal, N.L. (1992). Genetic influences on job
satisfaction: A reply to Cropanzano and James. Journal of Applied Psychology, 77, 89-93.
Bower, G.H. (198 1). Mood and memory. American psycho log is^. 36, 129-148.
Bower, G.H., & Mayer, J.D. (1985). Failure to replicate mood congruent retrieval. Bulletin of
fhe Psychonomic Sociefy, 23, 39-42.
Branscombe, N.R. (1985). Effects of hedonic valence and physiological arousal in emotion: A
comparison of two theoretical perspectives. Motivafion and Emotion, 9, 153-169.
Breckler, S.J., & Wiggins, E.C. (1989). Affect versusevaluation in the structure of attitudes. Journal
of Experimen fa1Social Psychology, 25, 253-27 I.
Brief, A.P., Butcher, A.B., & Roberson, L. (1995). Cookies, disposition, and job attitudes: The
effects of positive mood inducing events and negative affectivity on job satisfaction in a
field experiment. Organizational Behavior and Human Decision Processes, 62, 55-62.
Brief, A.P., & Roberson, L. (1989). Job attitude organization: An exploratory study. Journal of
Applied Social Psychology, 19, 7 17-727.
Buss, D.M., Gomes, M., Higgins, D.S., & Lauterbach, K. (1987). Tactics of manipulation. Journal
of Personalify and Social Psychology, 52, 1219-1229.
Cantor, J.R., Zillman, D., & Bryant, J. (1975). Enhancement of experienced sexual arousal in
response to erotic stimuli through misattribution of unrelated residual excitation. Journal
of Personalify and Social Psychology, 32, 69-75.
Carlson, M., & Miller, N. (1987). Explanation of the relation between negative mood and helping.
Psychological Bulletin, 102, 9 1- 108.
Carnevale, P.J.D., & Isen, A.M. (1986). The influence of positive affect and visual access on the
discovery of integrative solutions in bilateral negotiation. Organizafional Behavior and
Human Decision Process. 37, 1-1 3.
Clore, G.L. (1992). Cognitive phenomenology: Feelings and the construction ofjudgment. In L.L.
Martin & A. Tesser (Eds.), The Construction of social judgments (Vol. 10, pp. 133-163).
Hillsdale, NJ: Lawrence Erlbaum Associates.
Clore, G.L., Schwarz, N., & Conway, M. (1994). Affective causes and consequences of Social
Information Processing. In P.S. Wyer Jr. & T.K. Srull (Eds.), Handbook ofsocialcognition
(Vol. 1, pp. 323-417). Hillsdale, NJ: Lawrence Erlbaum Associates.
Conway, M., & Giannopoulos, C. (1993). Dysphoria and decision making: Limited information
use in the evaluation of multiattribute targets. Journal of Personality and Social
Psyc~hology,64, 6 13-623.
Cranny, C.J., Smith, P.C., & Stone, E.F. (1992). Job satisfaction: How people feel about their
jobs and how it aflects rheir performance. New York: Lexington Press.
Cropanzano, R., & James, K. (1990). Some methodological considerations for the behavioral
genetic analysis of work attitudes. Journal of Applied Psychology, 75,433-439.
Cropanzano, R., James, K., & Konovsky, M.A. (1993). Dispositional affectivity as a predictor
of work attitudes and job performance. Journal of Organizational Behavior, 14, 595-606.
Cunningham, M.R. (1979). Weather, mood, and helping behavior: Quasi-experiments with the
sunshine samaritan. Journal of Personality and Social Psychology, 37, 1947-1956.
Czajka, J . (1990). The relation of positive and negative affectivity to workplace attitudes. In L.R.
Jauch & J.L. Wall (Eds.), Academy of management bestpapers proceedings, 1990. Ada,
OH: Academy of Management.
Darwin, C. (1872[1965]). The Expression of emotions in man andanimals. Chicago, IL: University
of Chicago Press.
Davis-Blake, A., & Pfeffer, J. (1989). Just a mirage: The search for dispositional effects in
organizational research. Academy of Management Review, 14, 385-400.
Dawe.s, R.V. (1992). The structure of occupations: Beyond RIASEC. Journal of Vocational
Behavior, 40, 171-178.
Diener, E., Colvin, C.R., Pavot, W.G., & Allman, A. (1991). The psychic costs of intense positive
affect. Journal of Personality and Social Psychology, 61,492-503.
Diener, E., & Emmons, R.A. (1984). The independence of positive and negative affect. Journal
o f Personality and Social Psychology, 47, 1 105-I 1 17.
Diener, E., & Iran-Nejad, A. (1986). The relationship in experience between different types of
affect. Journal of Personality and Social Psychology, 50, 1 131-1 138.
Diener, E., Sandvick, E., & Pavot, W. (1990). Happiness is the frequency, not the intensity of
positive versus negative affect. In F. Strack, M. Argyle, & N. Schwarz (Eds.), Subjective
well-being: An interdisciplinary approach (Vol. 8, pp. 119-139). Oxford, England:
Pergamon Press.
Dipboye, R.L., Smith, C.S., & Howell, W.C. (1994). Understanding industrial and organizational
psychology: An integrated approach. Ft. W.orth, TX: Harcourt Brace College Publishers.
Dollard, J., Doob, L.W., Miller, N.E., Mowrer, O.H., & Sears, R.R. (1939). Frustration and
aggression. New Haven, CT: Yale University Press.
Eagly, A.H., & Chaiken, S. (1993). The Psychology of attitudes. Ft. Worth, TX: Harcourt Brace
Jovanovich College Publishers.
Edwards, K. (1990). The interplay of affect and cognition in attitude formation and change. Journal
of Personality and Social Psychology, 59, 202-2 16.
Eich, E. (1995). Searching for mood 'dependent memory. Psychological Science, 6,67-75.
Ekman, P. (1992). An argument for basic emotions. Cognition and Emotion, 6, 169-200.
Ellis, H.C., & Ashbrook, P.W. (1988). Resource allocation model of the effects of depressed mood
states on memory. In K. Fiedler & J . Forgas (Eds.), Affect, cognition, and social behavior
(pp. 25-43). Toronto: C.J. Hogrefe.
Fazio, R.H. (1986). How do attitudes guide behavior? In R.M. Sorrentino & E. T. Higgens (Eds.),
Handbook of motivation and cognition (pp. 204-243). New York: Guilford Press.
Finn, S.E. (1986). Stability of personality self-ratings over 30 years: Evidence for an agelcohort
interaction. Journal of Personality and Social Psychology, 50, 8 13-818.
Fisher, C. (1980). On the dubious wisdom of expecting job satisfaction to correlate with
performance. Academy of Management Review, 5, 607-6 12.
Fisher, C.D., & Locke, E.A. (1992). The new look in job satisfaction research and theory. In C.J.
Cranny, P.C. Smith, & E.F. Stone (Eds.), Job satisfaction: How people feel about their
jobs and how it affects theirperformance (Vol. I, pp. 165-194). New York: Lexington Press.
Fleming, I., Baum, A., & Weiss, L. (1987). Social density and perceived control as mediators of
crowding stress in high-density residential neighborhoods. Journal of Personality and
Social Psychology, 52, 899-906.
Folkman, S., & Lazarus, R.S. (1990). Coping and emotion. In N.L. Stein, B. Leventhal & T.
Trabasso (Eds.), Psychological and biological approaches to emotion (pp. 3 13-322).
Hillsdale, NJ: Erlbaum.
Freedman, J.L. (1975). Crowding and behavior. San Francisco, CA: Freeman.
Frijda, N.H. (1986). The emotions. Cambridge: Cambridge University Press.
Frijda, N.H. (1987). Emotion, cognitive structure, and action tendency. Cognition and Emotion,
I , 115-143.
Frijda, N. H. (1989). Aesthetic emotions and reality. American Psychologist, 44, 1546-1547.
Frijda, N.H. (1993). Moods, emotion episodes and emotions. In M. Lewis & J.M. Haviland (Eds.),
Handbook of emotions (pp. 38 1-403). New York: Guildford Press.
George, J. M. ( 1989). Mood and absence. Journal of Applied Psychology, 74, 3 17-324.
George, J.M. (1991). State or trait: Effects of positive mood on prosocial behaviors at work.
Journal of Applied Psychology, 76, 299-307.
George, J.M. (1992). The role of personality in organizational life: Issues and evidence. Journal
of Management, 18, 185-210.
George, J.M. & Brief, A.P. (1992). Feelingn good-doing good: A conceptual analysis of the mood
at work-organizational spontaneity relationship. Psychological Bulletin, 112, 3 10-329.
Gerhart, B. (1987). How important are dispositional factors as determinants of job satisfaction?
Implications for job design and other personnel programs. Journalof Applied Psychology,
72,366-373.
Gotlib, I.H., & Meyer, J.P. (1986). Factor analysis of the multiple affect adjective check list: A
separation of positive and negative affect. Journal of Personality and Social Psychology,
50, 1161-1165.
Green, D.P., Goldman, S.L., & Salovey, P. (1993). Measurement error masks bipolarity in affect
ratings. The Journal of Personality and Social Psychology, 64, 1029-1041.
Greenberg, J. (1982). Approaching equity and avoiding inequity in groups and organizations. In
J. Greenberg & R.L. Cohen (Eds.), Equity and justice in social behavior (pp. 389-435).
New York: Academic Press.
Griffitt, W., and Veitch, R. (1971). Hot and crowded: Influence of population density and
temperature on interpersonal affective behavior. Journal of Personality and Social
Psychology, 17, 92-98.
Hackett, R.D., & Guion, R.M. (1985). A reevaluation of the absenteeism-job satisfaction
relationship. Organizational Behavior and Human Decision Processes, 35, 340-38 1.
Hanisch, K.A., & Hulin, C.L. (1990). Job attitudes and organizational withdrawal: An
examination of retirement and other voluntary withdrawal behaviors. Journal of
Vocational Behavior, 37, 60-78.
Hastie, R., & Parke, B. (1986). The relation between memory and judgment depends on whether
the judgment task is memory-based or on-line. Psychological Review, 93,258-268.
Hersey, R.B. (1932). Worker's emotions in shop and home: A study of individual workers from
the psychological and physiological standpoint. Philadelphia: University of Pennsylvania
Press.
Herzberg, F., Mausner, B., & Snyderman, B.B. (1959). The motivation to work. New York: John
Wiley & Sons.
Hulin, C. (1991). Adaptation, persistence and commitment in organizations. In M.D. Dunnette
& L.M. Hough (Eds.), Handbook of industrial and organizational psychology (Vol. 25,
pp. 445-505). Palo Alto, CA: Consulting Psychologists Press.
laffaldano, M.T., & Muchinsky, P.M. (1985). Job satisfaction and job performance: A meta-
analysis. Psychological Bulletin, 97, 25 1-273.
llgen, D.R. (1971). Satisfaction with performance as a function of the initial level of expected
performance and the deviation from expectations. Organizational Behavior and Human
Performance. 6,345-36 1.
lsen, A.M. (1984). Toward understanding the role of affect in cognition. In R.S. Wyer, Jr. &
T.K. Srull (Eds.), Handbook of social cognition (Vol. 3, pp. 179-236). Hillsdale, NJ.:
Erlbaum.
Isen, A.M., & Baron, R.A. (199 1). Positive affect as a factor in organizational behavior. In L.L.
Cummings & B.M. Staw (Eds.), Research in organizalional behavior (Vol. 13, pp. 1-53).
Greenwich, CT: JAl Press.
Isen, A.M., & Daubman, K.A. (1984). The influence of affect on categorization. Journal of
Personalily and Social Psychology, 47, 1 206- 1 2 1 7.
Isen, A.M., Nygren, T.E., & Ashby, F.G. (1988). Influence of positive affect on the subjective
utility of gains and losses: It is just not worth the risk. Journal of Personalily and Social
Psychology, 55, 7 10-717.
Isen, A.M., & Patrick, R. (1983). The effect of positive feeling on risk-taking: When the chips
are down. Organizational Behasior and Human Peformance, 31, 194-202.
Izard, C.E. (1977). Human emotions. New York: Plenum.
Izard, C.E. (1992). Basic emotions, relations among emotions, and emotion-cognition relations.
Psychological Bullerin, 99, 56 1-565.
James, K., Clark, K., & Cropanzano, R. (1994). Positive and negative creativity in groups,
inslilurions, and organizations: A review and lheorelical exlension. Manuscript under
review.
James, W. ( 1884). What is an emotion? Mind, 9, 188-205.
James, W. (1890). Principles ofpsychology. New York: Holt.
Johnson, E., & Tversky, A. (1983). Affect, generalization, and the perception of risk. Journal of
Personalily and Social Psychology, 45, 20-3 1.
Johnson, J.A., & Ostendorf, F. (1993). Clarification of the five-factor model with the abridged
Big Five Dimensional Circumplex. Journal of Personalily and Social Psychology, 65,563-
576.
Judge, T.A. (1992). Dispositional perspective in human resources research. In G.R. Ferris & K.M.
Rowland (Eds.), Research in personality and human resources managemem (Vol. 10, pp.
31-72). Greenwich, CT: JAI Press.
Katzell, R.A. (1964). Personal values, job satisfaction, and job behavior. In H. Borow (Ed.), Man
in a world of work. Boston: Houghton Mifflin.
Kohn, M.L., & Schooler, C. (1982). Job conditions and personality: A longitudinal assessment
of their reciprocal effects. American Journal of Sociology, 87, 1257-1286.
Konovsky, M.A., Folger, R., & Cropanzano, R. (1987). Relative effects of procedural and
distributive justice on employee attitudes. Representative Research in Social Psychology,
17, 15-24.
Larsen, R.J., & Diener, E. (1987). Affect intensity as an individual difference characteristic: A
review. Journal of Research in Personality, 21, 1-3 1.
Larsen, R.J., & Diener, E. (1992). Promises and problems with the circumplex model of emotions.
Review of Personality and Social Psychology, 13, 25-29.
Larsen, R.J., & Kasimatis, M. (1990). Individual differences in entrainment of mood to the weekly
calendar. Journal of Personality and Social Psychology. 58, 164- 17 1.
Larsen, R.J., & Ketelaar, T. (1989). Extraversion, neuroticism, and susceptibility to positive and
negative mood induction procedures. Personality and Individual Dfferences, 10, 122 1-
1228.
Larsen, R.J., & Ketelaar, T. (1991). Personality and susceptibility to positive and negative
emotional states. The Journal of Personality and Social Psychology, 61, 132- 140.
Lawler, E.E. ( 1973). Motivation in work organizations. Monterey, CA: Brooks/ Cole Publishing.
Lazarus, R.S. (1991a). Emotion and adaptation. New York: Oxford University Press.
Lazarus, R.S. (1991b). Progress on a cognitive-motivational-relational theory of emotion.
American Psychologist. 46, 8 19-834.
Lazarus, R.S. (1991~).Cognition and motivation in emotion. American Psychologist, 46, 352-
367.
Levin, I., & Stokes, J.P. (1989). Dispositional approach to job satisfaction: Role of negative
affectivity. Journal of Applied Psychology, 74, 752-758.
Locke, E.A. ( 1976). The nature and causes ofjob satisfaction. In M.D. Dunnette (Ed.), Handbook
of industrial and organizational psychology. Chicago: Rand McNally.
Locke, E.A., & Latham, G.P. (1990). A theory of goal setling and task performance. Englewood
Cliffs, NJ: Prentice-Hall.
Lofquist, L.H., & Dawes, R.V. (1969). Adjustment to work: A psychological view of man's
problems in a ~clork-orientedsocielv. New York: Appleton-Century-Crofts, Educational
Division.
MacFarlane, J., Martin, C.L., & Williams, T.M. (1988). Mood fluctuations: Women versus men
and menstrual versus other cycles. Psychology o f Women Quarterly, 12, 201-223.
Manucia, G.K., Baumann, D.J., & Cialdini, R.B. (1984). Mood influences in helping: Direct effects
or side effects? Journal of Personality and Social Psychology, 46, 357-364.
Marco, C.A., & Suls, J . (1993). Daily stress and the trajectory of mood: Spillover, response
assimilation, contrast, and chronic negative affectivity. The Journal of Personality and
Social Psychology, 64, 1053- 1063.
Maslow, A. H. ( 1954). Motivation and personality. New York: Harper.
Mayer, J.D., & Gaschke, Y.N. (1988). The experience and meta-experience of mood. Journal of
Personalily and Social Psychology, 55, 102- 1 1 1.
Meyer, G.J., & Shack, J.R. (1989). Structural convergence of mood and personality: Evidence
for old and new directions. The Journal of Personality and Social Psychology, 57, 691-
706.
Miller, M.G., & Tesser, A. (1986). Effects of affective and cognitive focus on the attitude behavior
relation. Journal of Personality and Social Psychology, 51, 270-276.
Mitchell, T.R. (1974). Expectancy models of job satisfaction, occupational preferences and effort:
A theoretical, methodological, and empirical appraisal. Psychological Bulletin, 81, 1053-
1077.
Morris, W.N. (1989). Mood: Theframe of mind. New York: Springer-Verlag.
Oldham, G.R., & Fried, Y. (1987). Employee reactions to workspace characteristics. Journal of
Applied Psychology, 72, 75-80.
@tony, A., Clore, G.L., & Collins, A. (1988). The cognitive structure of emotions. Cambridge:
Cambridge University Press.
Ortony, A., & Turner, T.J. (1990). What's basic about basic emotions? Psychological Bulletin,
97, 3 15-331.
Osgood, C.E., Suci, G.J., & Tannenbaum, P.H. (1957). The measurement of meaning. Urbana:
University of lllinois Press.
Parkes, K.R. (1990). Coping, negative affectivity, and the work environment: Additive and
interactive predictors of mental health. Journal of Applied Psychology, 75, 399-408.
Persinger, M.A. (1975). Geophysical models for parapsychological experiences. Psychoenergetic

Systems, 1 , 63-74.
Plomin, R. (1990). The role of inheritance in behavior. Science, 248, 183-188.
Plutchik, R. (1994). The psychology and biology of emotion. New York: Harper Collins College
Publishers.
Podsakoff, P.M., & Williams, L.J. (1986). The relationship between job performance and job
satisfaction. In E.A. Locke (Ed.), Generalizing from laboratory tofield settings (pp. 207-
253). Lexington, MA: Lexington Press.
Porter, L.W. (1962). Job attitudes in management: 1. Perceived deficiencies in need fulfillment
as a function of job level. Journal oj'Applied Psychology, 46,375-384.
Rice, R. W., McFarlin, D.B., & Bennett, D.E. (1989). Standards of comparison and job satisfaction.
Journal of Applied Psychology, 74, 59 1-598.
Roseman, 1.J. (1984). Cognitivedeterminants of emotions: A structural theory. In P. Shaver (Ed.),
Review of personality and socialpsychology (Vol. 5 , pp. 11-36). Beverly Hills, CA: Sage.
Roseman, I.J., Spindel, M.S., & Jose, P.E. (1990). Appraisal of emotion eliciting events: Testing
a theory of discrete emotions. Journal of Personality and Social Psychology, 59, 99-915.
Rotton, J. (1983). Affective and cognitive consequences of malodorous pollution. Basic and
Applied Social Psychology, 4, 171- 191.
Russell, J.A. (1978). Evidence of convergent validity on the dimensions of affect., Journal of
Personality and Social Psychology, 36, 1152-1 168.
Russell, J.A. (1979). Affective space is bipolar. Journal of Personality and Social Psychology,
37, 345-356.
Russell, J.A. (1 980). A circumplex model of affect. Journal of Personality and Social Psychology,
39, 1161-1 178.
Russell, J.A., Lewicka, M., & Niit, T. (1989). A cross-cultural study of a circumplex model of
affect. The Journal of Personality and Social Psychology, 57, 848-856.
Russell, J.A., & Mehrabian, A. (1977). Evidence for a three-factor theory of emotion. Journal
of Research in Personality, 11, 273-294.
Russell, J.A., & Ridgeway, D. (1983). Dimensions underlying children's emotion concepts.
Developmental Psychology, 19, 795-804.
Salancik, G.R., & Pfeffer, J.R. (1977). An examination of need-satisfaction models ofjob attitudes.
Administrative Science Quarterly, 22,427-456.
Salancik, G.R., & Pfeffer, J.R. (1978). A social information processing approach to job attitudes.
Administrative Science Quarierly, 23, 224-252.
Scherer, K.R. (1984). On the nature and function of emotion: A component process approach.
In K.R. Scherer & P. Ekman (Eds.), Approaches to emotion (pp. 293-318). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Schwarz, N.,*& Clore, G.L. (1983). Mood, misattribution, and judgments of well-being:
Informative and directive functions of affective states. Journal of Personality and Social
Psychology, 45, 5 13-523.
Schwarz, N., & Clore, G.L. (1988). How do I feel about it? Informative functions of affective
states. In K. Fiedler & J . Forgas (Eds.), Affect, cognition, and social behavior (pp. 44-
62). Toronto: Hogrefe International.
Shaver, P., Schwartz, J., Kirson, D., & O'Connor, C. (1987). Emotion knowledge: Further
exploration of a prototype approach. Journal of Personality and Social Psychology, 52,
I06 1- 1086.
Sinclair, R.C. (1988). Mood, categorization breadth, and performance appraisal: The effects of
order of information acquisition and affective state on halo, accuracy, information retrieval,
and evaluations. Organizational Behavior and Human Decision Processes, 42, 22-46.
Sinclair, R.C., & Mark, M.M. (1992). The influence of mood state on judgment and action: Effects
on persuasion, categorization, social justice, person perception and judgmental accuracy.
In L.L. Martin & A. Tesser (Eds.), The consfrucfion of social judgmen~s(pp. 165-193).
Hillsdale, NJ: Lawrence Erlbaum.
Smith, C.A., & Ellsworth, P.C. (1985). Patterns of cognitive appraisal in emotion. Journal of
Personalify and Social Psychology, 48, 8 13-838.
Smith, C.A., & Ellsworth, P.C. (1987). Patterns of appraisal and emotion related to taking an
exam. Journal of Personalify and Social Psychology, 52, 475-488.
Smith, C.A., and Pope, L.K. (1992). Appraisal and Emotion: The interactional contributions of
dispositional and situational factors. In M.S. Clark (Ed.), Review of personalify andsocial
psychology, Vol. 14: Emofion and social behavior (pp. 32-62). Newbury Park, CA: Sage.
Staw, B.M., & Barsade, S.G. (1993). Affect and managerial performance: A test of the sadder-
but-wiser vs. happier-and-smarter hypotheses. Adminislralive Science Quarferly, 38, 304-
331.
Staw, B.M., & Ross, J. (1985). Stability in the midst of change: A dispositional approach to job
attitudes. Journal of Applied Psychology, 70, 469-480.
Stein, N.L., Trabasso, T., & Liwag, M. (1993). The representation and organization of emotional
experience: Unfolding the emotion episode. In M. Lewis and J.M. Haviland (Eds.),
Handbook of emofions (Vol. 13, pp. 279-300). New York: Guilford Press.
Stone, E.F. (1992). A critical analysis of Social Information Processing models ofjob perceptions
and job attitudes. In C.J. Cranny, P.C. Smith, & E.F. Stone (Eds.), Job safisfacfion:How
peoplefeel a b o u ~heir jobs and how if affecfs [heir performance (Vol. I, pp. 2 1-44). New
York: Lexington Press.
Taylor, S.E. (1991). The asymmetrical impact of positive and negative events: The mobilization-
minimization hypothesis. Psychological Bullefin, 110, 67-85.
Terkel, S. (1974). Working: People lalk a b o u ~whaf fhqv do all day and how they feel abouf whaf
fhe-vdo. New York: Pantheon Books.
Watson, D. (1988). lntraindividual and interindividual analyses of positive and negative affect:
Their relation to health complaints, perceived stress, and daily activities. Journal of
Personalify and Social Psyc~hology,54, 1020- 1030.
Watson, D., & Clark, L.A: (1984). Negative affectivity: This disposition to experience negative
emotional states. Ps.vc~hologic~a1 Bullefin, 96, 465-490.
Watson, D., & Clark, L.A. (199 1). Self-versus-peer ratings of specific emotional traits: Evidence
of convergent and discriminant validity. The Journal of Personali~yand Social Psychology,
60, 927-940.
Watson, D., Clark, L.A., & Carey, G. (1988). Positive and negative affectivity and their relation
to anxiety and depressive disorders. Journal of Abnormal Psychology, 97, 346-353.
Watson, D., Clark, L.A., & Tellegen, A. (1984). Cross-cultural convergence in the structure of
mood: A Japanese replication and a comparison with U.S. findings. Journal of Personalify
and Social Ps.vc.hology, 47, 127- 144.
Watson, D., Clark, L.A, & Tellegen, A. (1988). Development and validation of brief measures
of positive and negative affect: The PANAS scales. Journal of Personalify and Social
Psy~~hology, 54, 1063- 1070.
Watson, D., & Slack. A.K. (1993). General factors of affective temperament and their relation
to job satisfaction over time. Organiza~ionalBehavior and Human Decision Processes, 54,
181-202.
Watson, D., &Tellegen, A. ( 1985).Toward a consensual structure of mood. Psychological Bullefin,
98,2 19-235.
WeinertB. (1985). An attributional theory of achievement motivation and emotion. P.~,~~c~hological
Review, 92, 548-573.
Weiss, H.M. (1977). Subordinate imitation of supervisor behavior: The role or modeling in
organizational socialization. Organizational Behavior and Human Performance, 19, 89-
105.
Weiss, H.M., Nicholas, J.P., & Daus, C. (1993). Aflective and cognitive in/luenc'es on job
satisfaction. San Francisco, CA: Society of Industrial/Organizational Psychology.
Weiss, H.M., & Shaw, J.B. (1979). Social influences on judgments about tasks. Organizational
Behavior and Human Perjbrmancae, 24, 126- 140.
Weitz, J. (1952). A neglected concept in the study of job satisfaction. Personnel Psychology. 5,
20 1-205.
Werner, P.D., & Pervin, L.A. (1986). The content of personality items. The ~ournalof ~ersonalit~
and Social Psycahology, 51, 622-628.
Wharton, A.S., & Erickson, R.J. (1993). Managing emotions on the job and at home:
Understanding the consequences of multiple emotional roles. Academy of Management
Review. 18, 457-486.
Willerman, L. (1979). The psychology oj'individual and group differences. San Francisco: W.H.
Freeman.
Wilson, T.D., & Hodges, S.D. (1992). Attitudes as temporary constructions. In L. Martin & A.
Tesser (Eds.), The construction of social judgments (Vol. 10, pp. 37-65). Hillsdale, NJ:
Lawrence Erlbaum Associates.
Worth, L.T., & Mackie, D.M. (1987). Cognitive mediation of positive mood in persuasion. Social
Cognition, 5, 76-94.
Wright, W.F., & Bower, G .H. (1992). Mood effects on subjective probability assessment.
Organizational Behavior and Human Decision Processes, 52, 276-29 1.
Zalesny, M.D., & Ford, -1.K. (1990). Extending the social information processing perspective: New
links to attitudes, behaviors, and perceptions. Organizational Behavior and Human
Decision Processes. 47, 205-246.
Zevon, M.A., & Tellegen, A. (1982). The structure of mood change: An idiographiclnomothetic
analysis. The Journal of Personality and Social Psychology, 43, 1 1 1-122.
Zillman, D. (1979). Hostility and aggression. Hillsdale, NJ: Lawrence Erlbaum Associates.
Zillman, D., & Bryant, J. (1974). The effect of residual excitation on the emotional response to
provocation and delayed aggressive behavior. Journal o j Personality and Social
Psychology. 30, 782-79 I.
View publication stats

Affective Events Theory A Theoretical Discussion o PDF

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Affective Events Theory A Theoretical Discussion o PDF

Caricato da

Copyright:

Formati disponibili

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Affective Events Theory: A Theoretical Discussion of The Structure, Cause and

Article in Research in Organizational Behavior · January 1996

The user has requested enhancement of the downloaded file.

Howard M. Weiss and Russell Cropanzano

Research in Organizational Behavior, Volume 18, pages 1-74.

THREE THEORETICAL APPROACHES

Cognitive Judgment Approach

By far, the cognitive judgment approach has dominated the theoretical

the work environment is represented as a set of concrete or abstract features

Figure 1. Cognitive Judgment Approach

Social Influences Approach

Figure 2. Cognitive Judgment and Social Influences Approaches

However, we also believe that there is a genuine source of conflict between

The basic idea of the Dispositional approach is this: to some degree, a

Personality Determinants of Dispositional Satisfaction

variance in satisfaction is not by itself very interesting. Accounting for more

Genetic influences on Job Satisfaction

AFFECTIVE EVENTS THEORY: AN OVERVIEW

In contrasting Affective Events Theory with traditional theories the first

In Figure 3 we have presented a tentative macro structure to the affective events

PREVIOUS RESEARCH CONSISTENT WITH AET

In 1932 Rexford Hersey published his research on Workers' Emotions in

as positive and negative, and document daily fluctuations in affect. Second,

Herzberg, Mausner, a n d Snyderman (1959)

Herzberg maintained that satisfaction and dissatisfaction were not poles of

THE NATURE OF EMOTIONS AND MOODS

be followed, in turn, by discussions of the way events come to elicit affective

Affective Events Theory 19

I The Stucture of Emotions

of structure but, instead, a surfeit of perspectives, points-of-view, and

Disgust Alarm Cheerhlness Affection Disappointment

Source: Adapted from Shaver, Schwartz, Kirson, and O'Conner (1987).

Figure 5. Semantic Organization of Emotion Words

lrnplications for Affective Events Theory

The Structure of Mood

So far we have been discussing mood as if it were a single thing. However,

Positive and Negative Affectivity

Although still limited, cross-cultural investigations have also produced

Hedonic Tone and Intensity

W h y the Confusion Over Factor Structures?

At first glance the two previous sections appear to completely contradict

Robustness Across Multiple Operations

As we have seen, both the NA/PA and the Hedonic Tone/Intensity

indicative of high positive affect (such as "energetic") or high negative affect

The Issue of Error Variance

A series of studies by Green, Goldman, and Salovey (1993) is also

The Issue of Time Frame

integration .With Other Research

THE EFFECTS OF EVENTS ON EMOTION AND MOOD

Initial or primary appraisal appears based on "concern relevance" (Frijda,

Second, in agreement with most cognitive appraisal theorists (Lazarus,

All cognitive appraisal theorists argue that primary appraisal is followed by

Mood Generating Events

There exists quite a substantial empirical literature which speaks to the

Source: Adapted from Stein, Trabassb. and Liwag (1993).

Figure 7. Appraisal Flow for Happiness and Sadness

have demonstrated the efficacy of manipulating mood by showing brief videos,

OTHER CAUSES OF AFFECT LEVELS AT WORK

Despite this, we were careful to note that these observations do nothing to

Main Effects of Affective Dispositions

Research suggests that affective dispositions exert main effects on job

we see evidence that affective traits serve as predispositions to respond within